ComfyUI version v0.3.66

This reverts commit b1467da480.

README.md

@@ -197,7 +197,9 @@ comfy install
 
 ## Manual Install (Windows, Linux)
 
-Python 3.13 is very well supported. If you have trouble with some custom node dependencies you can try 3.12
+Python 3.14 will work if you comment out the `kornia` dependency in the requirements.txt file (breaks the canny node) but it is not recommended.
+
+Python 3.13 is very well supported. If you have trouble with some custom node dependencies on 3.13 you can try 3.12
 
 Git clone this repo.
 
@@ -253,7 +255,7 @@ This is the command to install the Pytorch xpu nightly which might have some per
 
 Nvidia users should install stable pytorch using this command:
 
-```pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu129```
+```pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu130```
 
 This is the command to install pytorch nightly instead which might have performance improvements.
 

comfy/ldm/chroma_radiance/model.py

@@ -189,15 +189,15 @@ class ChromaRadiance(Chroma):
         nerf_pixels = nn.functional.unfold(img_orig, kernel_size=patch_size, stride=patch_size)
         nerf_pixels = nerf_pixels.transpose(1, 2) # -> [B, NumPatches, C * P * P]
 
+        # Reshape for per-patch processing
+        nerf_hidden = img_out.reshape(B * num_patches, params.hidden_size)
+        nerf_pixels = nerf_pixels.reshape(B * num_patches, C, patch_size**2).transpose(1, 2)
+
         if params.nerf_tile_size > 0 and num_patches > params.nerf_tile_size:
             # Enable tiling if nerf_tile_size isn't 0 and we actually have more patches than
             # the tile size.
-            img_dct = self.forward_tiled_nerf(img_out, nerf_pixels, B, C, num_patches, patch_size, params)
+            img_dct = self.forward_tiled_nerf(nerf_hidden, nerf_pixels, B, C, num_patches, patch_size, params)
         else:
-            # Reshape for per-patch processing
-            nerf_hidden = img_out.reshape(B * num_patches, params.hidden_size)
-            nerf_pixels = nerf_pixels.reshape(B * num_patches, C, patch_size**2).transpose(1, 2)
-
             # Get DCT-encoded pixel embeddings [pixel-dct]
             img_dct = self.nerf_image_embedder(nerf_pixels)
 
@@ -240,17 +240,8 @@ class ChromaRadiance(Chroma):
             end = min(i + tile_size, num_patches)
 
             # Slice the current tile from the input tensors
-            nerf_hidden_tile = nerf_hidden[:, i:end, :]
-            nerf_pixels_tile = nerf_pixels[:, i:end, :]
-
-            # Get the actual number of patches in this tile (can be smaller for the last tile)
-            num_patches_tile = nerf_hidden_tile.shape[1]
-
-            # Reshape the tile for per-patch processing
-            # [B, NumPatches_tile, D] -> [B * NumPatches_tile, D]
-            nerf_hidden_tile = nerf_hidden_tile.reshape(batch * num_patches_tile, params.hidden_size)
-            # [B, NumPatches_tile, C*P*P] -> [B*NumPatches_tile, C, P*P] -> [B*NumPatches_tile, P*P, C]
-            nerf_pixels_tile = nerf_pixels_tile.reshape(batch * num_patches_tile, channels, patch_size**2).transpose(1, 2)
+            nerf_hidden_tile = nerf_hidden[i * batch:end * batch]
+            nerf_pixels_tile = nerf_pixels[i * batch:end * batch]
 
             # get DCT-encoded pixel embeddings [pixel-dct]
             img_dct_tile = self.nerf_image_embedder(nerf_pixels_tile)

comfy/ldm/mmaudio/vae/activations.py

@@ -0,0 +1,120 @@
+# Implementation adapted from https://github.com/EdwardDixon/snake under the MIT license.
+#   LICENSE is in incl_licenses directory.
+
+import torch
+from torch import nn, sin, pow
+from torch.nn import Parameter
+import comfy.model_management
+
+class Snake(nn.Module):
+    '''
+    Implementation of a sine-based periodic activation function
+    Shape:
+        - Input: (B, C, T)
+        - Output: (B, C, T), same shape as the input
+    Parameters:
+        - alpha - trainable parameter
+    References:
+        - This activation function is from this paper by Liu Ziyin, Tilman Hartwig, Masahito Ueda:
+        https://arxiv.org/abs/2006.08195
+    Examples:
+        >>> a1 = snake(256)
+        >>> x = torch.randn(256)
+        >>> x = a1(x)
+    '''
+    def __init__(self, in_features, alpha=1.0, alpha_trainable=True, alpha_logscale=False):
+        '''
+        Initialization.
+        INPUT:
+            - in_features: shape of the input
+            - alpha: trainable parameter
+            alpha is initialized to 1 by default, higher values = higher-frequency.
+            alpha will be trained along with the rest of your model.
+        '''
+        super(Snake, self).__init__()
+        self.in_features = in_features
+
+        # initialize alpha
+        self.alpha_logscale = alpha_logscale
+        if self.alpha_logscale:
+            self.alpha = Parameter(torch.empty(in_features))
+        else:
+            self.alpha = Parameter(torch.empty(in_features))
+
+        self.alpha.requires_grad = alpha_trainable
+
+        self.no_div_by_zero = 0.000000001
+
+    def forward(self, x):
+        '''
+        Forward pass of the function.
+        Applies the function to the input elementwise.
+        Snake ∶= x + 1/a * sin^2 (xa)
+        '''
+        alpha = comfy.model_management.cast_to(self.alpha, dtype=x.dtype, device=x.device).unsqueeze(0).unsqueeze(-1) # line up with x to [B, C, T]
+        if self.alpha_logscale:
+            alpha = torch.exp(alpha)
+        x = x + (1.0 / (alpha + self.no_div_by_zero)) * pow(sin(x * alpha), 2)
+
+        return x
+
+
+class SnakeBeta(nn.Module):
+    '''
+    A modified Snake function which uses separate parameters for the magnitude of the periodic components
+    Shape:
+        - Input: (B, C, T)
+        - Output: (B, C, T), same shape as the input
+    Parameters:
+        - alpha - trainable parameter that controls frequency
+        - beta - trainable parameter that controls magnitude
+    References:
+        - This activation function is a modified version based on this paper by Liu Ziyin, Tilman Hartwig, Masahito Ueda:
+        https://arxiv.org/abs/2006.08195
+    Examples:
+        >>> a1 = snakebeta(256)
+        >>> x = torch.randn(256)
+        >>> x = a1(x)
+    '''
+    def __init__(self, in_features, alpha=1.0, alpha_trainable=True, alpha_logscale=False):
+        '''
+        Initialization.
+        INPUT:
+            - in_features: shape of the input
+            - alpha - trainable parameter that controls frequency
+            - beta - trainable parameter that controls magnitude
+            alpha is initialized to 1 by default, higher values = higher-frequency.
+            beta is initialized to 1 by default, higher values = higher-magnitude.
+            alpha will be trained along with the rest of your model.
+        '''
+        super(SnakeBeta, self).__init__()
+        self.in_features = in_features
+
+        # initialize alpha
+        self.alpha_logscale = alpha_logscale
+        if self.alpha_logscale:
+            self.alpha = Parameter(torch.empty(in_features))
+            self.beta = Parameter(torch.empty(in_features))
+        else:
+            self.alpha = Parameter(torch.empty(in_features))
+            self.beta = Parameter(torch.empty(in_features))
+
+        self.alpha.requires_grad = alpha_trainable
+        self.beta.requires_grad = alpha_trainable
+
+        self.no_div_by_zero = 0.000000001
+
+    def forward(self, x):
+        '''
+        Forward pass of the function.
+        Applies the function to the input elementwise.
+        SnakeBeta ∶= x + 1/b * sin^2 (xa)
+        '''
+        alpha = comfy.model_management.cast_to(self.alpha, dtype=x.dtype, device=x.device).unsqueeze(0).unsqueeze(-1) # line up with x to [B, C, T]
+        beta = comfy.model_management.cast_to(self.beta, dtype=x.dtype, device=x.device).unsqueeze(0).unsqueeze(-1)
+        if self.alpha_logscale:
+            alpha = torch.exp(alpha)
+            beta = torch.exp(beta)
+        x = x + (1.0 / (beta + self.no_div_by_zero)) * pow(sin(x * alpha), 2)
+
+        return x

comfy/ldm/mmaudio/vae/alias_free_torch.py

@@ -0,0 +1,157 @@
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import math
+import comfy.model_management
+
+if 'sinc' in dir(torch):
+    sinc = torch.sinc
+else:
+    # This code is adopted from adefossez's julius.core.sinc under the MIT License
+    # https://adefossez.github.io/julius/julius/core.html
+    #   LICENSE is in incl_licenses directory.
+    def sinc(x: torch.Tensor):
+        """
+        Implementation of sinc, i.e. sin(pi * x) / (pi * x)
+        __Warning__: Different to julius.sinc, the input is multiplied by `pi`!
+        """
+        return torch.where(x == 0,
+                           torch.tensor(1., device=x.device, dtype=x.dtype),
+                           torch.sin(math.pi * x) / math.pi / x)
+
+
+# This code is adopted from adefossez's julius.lowpass.LowPassFilters under the MIT License
+# https://adefossez.github.io/julius/julius/lowpass.html
+#   LICENSE is in incl_licenses directory.
+def kaiser_sinc_filter1d(cutoff, half_width, kernel_size): # return filter [1,1,kernel_size]
+    even = (kernel_size % 2 == 0)
+    half_size = kernel_size // 2
+
+    #For kaiser window
+    delta_f = 4 * half_width
+    A = 2.285 * (half_size - 1) * math.pi * delta_f + 7.95
+    if A > 50.:
+        beta = 0.1102 * (A - 8.7)
+    elif A >= 21.:
+        beta = 0.5842 * (A - 21)**0.4 + 0.07886 * (A - 21.)
+    else:
+        beta = 0.
+    window = torch.kaiser_window(kernel_size, beta=beta, periodic=False)
+
+    # ratio = 0.5/cutoff -> 2 * cutoff = 1 / ratio
+    if even:
+        time = (torch.arange(-half_size, half_size) + 0.5)
+    else:
+        time = torch.arange(kernel_size) - half_size
+    if cutoff == 0:
+        filter_ = torch.zeros_like(time)
+    else:
+        filter_ = 2 * cutoff * window * sinc(2 * cutoff * time)
+        # Normalize filter to have sum = 1, otherwise we will have a small leakage
+        # of the constant component in the input signal.
+        filter_ /= filter_.sum()
+        filter = filter_.view(1, 1, kernel_size)
+
+    return filter
+
+
+class LowPassFilter1d(nn.Module):
+    def __init__(self,
+                 cutoff=0.5,
+                 half_width=0.6,
+                 stride: int = 1,
+                 padding: bool = True,
+                 padding_mode: str = 'replicate',
+                 kernel_size: int = 12):
+        # kernel_size should be even number for stylegan3 setup,
+        # in this implementation, odd number is also possible.
+        super().__init__()
+        if cutoff < -0.:
+            raise ValueError("Minimum cutoff must be larger than zero.")
+        if cutoff > 0.5:
+            raise ValueError("A cutoff above 0.5 does not make sense.")
+        self.kernel_size = kernel_size
+        self.even = (kernel_size % 2 == 0)
+        self.pad_left = kernel_size // 2 - int(self.even)
+        self.pad_right = kernel_size // 2
+        self.stride = stride
+        self.padding = padding
+        self.padding_mode = padding_mode
+        filter = kaiser_sinc_filter1d(cutoff, half_width, kernel_size)
+        self.register_buffer("filter", filter)
+
+    #input [B, C, T]
+    def forward(self, x):
+        _, C, _ = x.shape
+
+        if self.padding:
+            x = F.pad(x, (self.pad_left, self.pad_right),
+                      mode=self.padding_mode)
+        out = F.conv1d(x, comfy.model_management.cast_to(self.filter.expand(C, -1, -1), dtype=x.dtype, device=x.device),
+                       stride=self.stride, groups=C)
+
+        return out
+
+
+class UpSample1d(nn.Module):
+    def __init__(self, ratio=2, kernel_size=None):
+        super().__init__()
+        self.ratio = ratio
+        self.kernel_size = int(6 * ratio // 2) * 2 if kernel_size is None else kernel_size
+        self.stride = ratio
+        self.pad = self.kernel_size // ratio - 1
+        self.pad_left = self.pad * self.stride + (self.kernel_size - self.stride) // 2
+        self.pad_right = self.pad * self.stride + (self.kernel_size - self.stride + 1) // 2
+        filter = kaiser_sinc_filter1d(cutoff=0.5 / ratio,
+                                      half_width=0.6 / ratio,
+                                      kernel_size=self.kernel_size)
+        self.register_buffer("filter", filter)
+
+    # x: [B, C, T]
+    def forward(self, x):
+        _, C, _ = x.shape
+
+        x = F.pad(x, (self.pad, self.pad), mode='replicate')
+        x = self.ratio * F.conv_transpose1d(
+            x, comfy.model_management.cast_to(self.filter.expand(C, -1, -1), dtype=x.dtype, device=x.device), stride=self.stride, groups=C)
+        x = x[..., self.pad_left:-self.pad_right]
+
+        return x
+
+
+class DownSample1d(nn.Module):
+    def __init__(self, ratio=2, kernel_size=None):
+        super().__init__()
+        self.ratio = ratio
+        self.kernel_size = int(6 * ratio // 2) * 2 if kernel_size is None else kernel_size
+        self.lowpass = LowPassFilter1d(cutoff=0.5 / ratio,
+                                       half_width=0.6 / ratio,
+                                       stride=ratio,
+                                       kernel_size=self.kernel_size)
+
+    def forward(self, x):
+        xx = self.lowpass(x)
+
+        return xx
+
+class Activation1d(nn.Module):
+    def __init__(self,
+                 activation,
+                 up_ratio: int = 2,
+                 down_ratio: int = 2,
+                 up_kernel_size: int = 12,
+                 down_kernel_size: int = 12):
+        super().__init__()
+        self.up_ratio = up_ratio
+        self.down_ratio = down_ratio
+        self.act = activation
+        self.upsample = UpSample1d(up_ratio, up_kernel_size)
+        self.downsample = DownSample1d(down_ratio, down_kernel_size)
+
+    # x: [B,C,T]
+    def forward(self, x):
+        x = self.upsample(x)
+        x = self.act(x)
+        x = self.downsample(x)
+
+        return x

comfy/ldm/mmaudio/vae/autoencoder.py

@@ -0,0 +1,156 @@
+from typing import Literal
+
+import torch
+import torch.nn as nn
+
+from .distributions import DiagonalGaussianDistribution
+from .vae import VAE_16k
+from .bigvgan import BigVGANVocoder
+import logging
+
+try:
+    import torchaudio
+except:
+    logging.warning("torchaudio missing, MMAudio VAE model will be broken")
+
+def dynamic_range_compression_torch(x, C=1, clip_val=1e-5, *, norm_fn):
+    return norm_fn(torch.clamp(x, min=clip_val) * C)
+
+
+def spectral_normalize_torch(magnitudes, norm_fn):
+    output = dynamic_range_compression_torch(magnitudes, norm_fn=norm_fn)
+    return output
+
+class MelConverter(nn.Module):
+
+    def __init__(
+        self,
+        *,
+        sampling_rate: float,
+        n_fft: int,
+        num_mels: int,
+        hop_size: int,
+        win_size: int,
+        fmin: float,
+        fmax: float,
+        norm_fn,
+    ):
+        super().__init__()
+        self.sampling_rate = sampling_rate
+        self.n_fft = n_fft
+        self.num_mels = num_mels
+        self.hop_size = hop_size
+        self.win_size = win_size
+        self.fmin = fmin
+        self.fmax = fmax
+        self.norm_fn = norm_fn
+
+        # mel = librosa_mel_fn(sr=self.sampling_rate,
+        #                      n_fft=self.n_fft,
+        #                      n_mels=self.num_mels,
+        #                      fmin=self.fmin,
+        #                      fmax=self.fmax)
+        # mel_basis = torch.from_numpy(mel).float()
+        mel_basis = torch.empty((num_mels, 1 + n_fft // 2))
+        hann_window = torch.hann_window(self.win_size)
+
+        self.register_buffer('mel_basis', mel_basis)
+        self.register_buffer('hann_window', hann_window)
+
+    @property
+    def device(self):
+        return self.mel_basis.device
+
+    def forward(self, waveform: torch.Tensor, center: bool = False) -> torch.Tensor:
+        waveform = waveform.clamp(min=-1., max=1.).to(self.device)
+
+        waveform = torch.nn.functional.pad(
+            waveform.unsqueeze(1),
+            [int((self.n_fft - self.hop_size) / 2),
+             int((self.n_fft - self.hop_size) / 2)],
+            mode='reflect')
+        waveform = waveform.squeeze(1)
+
+        spec = torch.stft(waveform,
+                          self.n_fft,
+                          hop_length=self.hop_size,
+                          win_length=self.win_size,
+                          window=self.hann_window,
+                          center=center,
+                          pad_mode='reflect',
+                          normalized=False,
+                          onesided=True,
+                          return_complex=True)
+
+        spec = torch.view_as_real(spec)
+        spec = torch.sqrt(spec.pow(2).sum(-1) + (1e-9))
+        spec = torch.matmul(self.mel_basis, spec)
+        spec = spectral_normalize_torch(spec, self.norm_fn)
+
+        return spec
+
+class AudioAutoencoder(nn.Module):
+
+    def __init__(
+        self,
+        *,
+        # ckpt_path: str,
+        mode=Literal['16k', '44k'],
+        need_vae_encoder: bool = True,
+    ):
+        super().__init__()
+
+        assert mode == "16k", "Only 16k mode is supported currently."
+        self.mel_converter = MelConverter(sampling_rate=16_000,
+                            n_fft=1024,
+                            num_mels=80,
+                            hop_size=256,
+                            win_size=1024,
+                            fmin=0,
+                            fmax=8_000,
+                            norm_fn=torch.log10)
+
+        self.vae = VAE_16k().eval()
+
+        bigvgan_config = {
+            "resblock": "1",
+            "num_mels": 80,
+            "upsample_rates": [4, 4, 2, 2, 2, 2],
+            "upsample_kernel_sizes": [8, 8, 4, 4, 4, 4],
+            "upsample_initial_channel": 1536,
+            "resblock_kernel_sizes": [3, 7, 11],
+            "resblock_dilation_sizes": [
+                [1, 3, 5],
+                [1, 3, 5],
+                [1, 3, 5],
+            ],
+            "activation": "snakebeta",
+            "snake_logscale": True,
+        }
+
+        self.vocoder = BigVGANVocoder(
+            bigvgan_config
+        ).eval()
+
+    @torch.inference_mode()
+    def encode_audio(self, x) -> DiagonalGaussianDistribution:
+        # x: (B * L)
+        mel = self.mel_converter(x)
+        dist = self.vae.encode(mel)
+
+        return dist
+
+    @torch.no_grad()
+    def decode(self, z):
+        mel_decoded = self.vae.decode(z)
+        audio = self.vocoder(mel_decoded)
+
+        audio = torchaudio.functional.resample(audio, 16000, 44100)
+        return audio
+
+    @torch.no_grad()
+    def encode(self, audio):
+        audio = audio.mean(dim=1)
+        audio = torchaudio.functional.resample(audio, 44100, 16000)
+        dist = self.encode_audio(audio)
+        return dist.mean

comfy/ldm/mmaudio/vae/bigvgan.py

@@ -0,0 +1,219 @@
+# Copyright (c) 2022 NVIDIA CORPORATION.
+#   Licensed under the MIT license.
+
+# Adapted from https://github.com/jik876/hifi-gan under the MIT license.
+#   LICENSE is in incl_licenses directory.
+
+import torch
+import torch.nn as nn
+from types import SimpleNamespace
+from . import activations
+from .alias_free_torch import Activation1d
+import comfy.ops
+ops = comfy.ops.disable_weight_init
+
+def get_padding(kernel_size, dilation=1):
+    return int((kernel_size * dilation - dilation) / 2)
+
+class AMPBlock1(torch.nn.Module):
+
+    def __init__(self, h, channels, kernel_size=3, dilation=(1, 3, 5), activation=None):
+        super(AMPBlock1, self).__init__()
+        self.h = h
+
+        self.convs1 = nn.ModuleList([
+                ops.Conv1d(channels,
+                       channels,
+                       kernel_size,
+                       1,
+                       dilation=dilation[0],
+                       padding=get_padding(kernel_size, dilation[0])),
+                ops.Conv1d(channels,
+                       channels,
+                       kernel_size,
+                       1,
+                       dilation=dilation[1],
+                       padding=get_padding(kernel_size, dilation[1])),
+                ops.Conv1d(channels,
+                       channels,
+                       kernel_size,
+                       1,
+                       dilation=dilation[2],
+                       padding=get_padding(kernel_size, dilation[2]))
+        ])
+
+        self.convs2 = nn.ModuleList([
+                ops.Conv1d(channels,
+                       channels,
+                       kernel_size,
+                       1,
+                       dilation=1,
+                       padding=get_padding(kernel_size, 1)),
+                ops.Conv1d(channels,
+                       channels,
+                       kernel_size,
+                       1,
+                       dilation=1,
+                       padding=get_padding(kernel_size, 1)),
+                ops.Conv1d(channels,
+                       channels,
+                       kernel_size,
+                       1,
+                       dilation=1,
+                       padding=get_padding(kernel_size, 1))
+        ])
+
+        self.num_layers = len(self.convs1) + len(self.convs2)  # total number of conv layers
+
+        if activation == 'snake':  # periodic nonlinearity with snake function and anti-aliasing
+            self.activations = nn.ModuleList([
+                Activation1d(
+                    activation=activations.Snake(channels, alpha_logscale=h.snake_logscale))
+                for _ in range(self.num_layers)
+            ])
+        elif activation == 'snakebeta':  # periodic nonlinearity with snakebeta function and anti-aliasing
+            self.activations = nn.ModuleList([
+                Activation1d(
+                    activation=activations.SnakeBeta(channels, alpha_logscale=h.snake_logscale))
+                for _ in range(self.num_layers)
+            ])
+        else:
+            raise NotImplementedError(
+                "activation incorrectly specified. check the config file and look for 'activation'."
+            )
+
+    def forward(self, x):
+        acts1, acts2 = self.activations[::2], self.activations[1::2]
+        for c1, c2, a1, a2 in zip(self.convs1, self.convs2, acts1, acts2):
+            xt = a1(x)
+            xt = c1(xt)
+            xt = a2(xt)
+            xt = c2(xt)
+            x = xt + x
+
+        return x
+
+
+class AMPBlock2(torch.nn.Module):
+
+    def __init__(self, h, channels, kernel_size=3, dilation=(1, 3), activation=None):
+        super(AMPBlock2, self).__init__()
+        self.h = h
+
+        self.convs = nn.ModuleList([
+                ops.Conv1d(channels,
+                       channels,
+                       kernel_size,
+                       1,
+                       dilation=dilation[0],
+                       padding=get_padding(kernel_size, dilation[0])),
+                ops.Conv1d(channels,
+                       channels,
+                       kernel_size,
+                       1,
+                       dilation=dilation[1],
+                       padding=get_padding(kernel_size, dilation[1]))
+        ])
+
+        self.num_layers = len(self.convs)  # total number of conv layers
+
+        if activation == 'snake':  # periodic nonlinearity with snake function and anti-aliasing
+            self.activations = nn.ModuleList([
+                Activation1d(
+                    activation=activations.Snake(channels, alpha_logscale=h.snake_logscale))
+                for _ in range(self.num_layers)
+            ])
+        elif activation == 'snakebeta':  # periodic nonlinearity with snakebeta function and anti-aliasing
+            self.activations = nn.ModuleList([
+                Activation1d(
+                    activation=activations.SnakeBeta(channels, alpha_logscale=h.snake_logscale))
+                for _ in range(self.num_layers)
+            ])
+        else:
+            raise NotImplementedError(
+                "activation incorrectly specified. check the config file and look for 'activation'."
+            )
+
+    def forward(self, x):
+        for c, a in zip(self.convs, self.activations):
+            xt = a(x)
+            xt = c(xt)
+            x = xt + x
+
+        return x
+
+
+class BigVGANVocoder(torch.nn.Module):
+    # this is our main BigVGAN model. Applies anti-aliased periodic activation for resblocks.
+    def __init__(self, h):
+        super().__init__()
+        if isinstance(h, dict):
+            h = SimpleNamespace(**h)
+        self.h = h
+
+        self.num_kernels = len(h.resblock_kernel_sizes)
+        self.num_upsamples = len(h.upsample_rates)
+
+        # pre conv
+        self.conv_pre = ops.Conv1d(h.num_mels, h.upsample_initial_channel, 7, 1, padding=3)
+
+        # define which AMPBlock to use. BigVGAN uses AMPBlock1 as default
+        resblock = AMPBlock1 if h.resblock == '1' else AMPBlock2
+
+        # transposed conv-based upsamplers. does not apply anti-aliasing
+        self.ups = nn.ModuleList()
+        for i, (u, k) in enumerate(zip(h.upsample_rates, h.upsample_kernel_sizes)):
+            self.ups.append(
+                nn.ModuleList([
+                        ops.ConvTranspose1d(h.upsample_initial_channel // (2**i),
+                                        h.upsample_initial_channel // (2**(i + 1)),
+                                        k,
+                                        u,
+                                        padding=(k - u) // 2)
+                ]))
+
+        # residual blocks using anti-aliased multi-periodicity composition modules (AMP)
+        self.resblocks = nn.ModuleList()
+        for i in range(len(self.ups)):
+            ch = h.upsample_initial_channel // (2**(i + 1))
+            for j, (k, d) in enumerate(zip(h.resblock_kernel_sizes, h.resblock_dilation_sizes)):
+                self.resblocks.append(resblock(h, ch, k, d, activation=h.activation))
+
+        # post conv
+        if h.activation == "snake":  # periodic nonlinearity with snake function and anti-aliasing
+            activation_post = activations.Snake(ch, alpha_logscale=h.snake_logscale)
+            self.activation_post = Activation1d(activation=activation_post)
+        elif h.activation == "snakebeta":  # periodic nonlinearity with snakebeta function and anti-aliasing
+            activation_post = activations.SnakeBeta(ch, alpha_logscale=h.snake_logscale)
+            self.activation_post = Activation1d(activation=activation_post)
+        else:
+            raise NotImplementedError(
+                "activation incorrectly specified. check the config file and look for 'activation'."
+            )
+
+        self.conv_post = ops.Conv1d(ch, 1, 7, 1, padding=3)
+
+
+    def forward(self, x):
+        # pre conv
+        x = self.conv_pre(x)
+
+        for i in range(self.num_upsamples):
+            # upsampling
+            for i_up in range(len(self.ups[i])):
+                x = self.ups[i][i_up](x)
+            # AMP blocks
+            xs = None
+            for j in range(self.num_kernels):
+                if xs is None:
+                    xs = self.resblocks[i * self.num_kernels + j](x)
+                else:
+                    xs += self.resblocks[i * self.num_kernels + j](x)
+            x = xs / self.num_kernels
+
+        # post conv
+        x = self.activation_post(x)
+        x = self.conv_post(x)
+        x = torch.tanh(x)
+
+        return x

comfy/ldm/mmaudio/vae/distributions.py

@@ -0,0 +1,92 @@
+import torch
+import numpy as np
+
+
+class AbstractDistribution:
+    def sample(self):
+        raise NotImplementedError()
+
+    def mode(self):
+        raise NotImplementedError()
+
+
+class DiracDistribution(AbstractDistribution):
+    def __init__(self, value):
+        self.value = value
+
+    def sample(self):
+        return self.value
+
+    def mode(self):
+        return self.value
+
+
+class DiagonalGaussianDistribution(object):
+    def __init__(self, parameters, deterministic=False):
+        self.parameters = parameters
+        self.mean, self.logvar = torch.chunk(parameters, 2, dim=1)
+        self.logvar = torch.clamp(self.logvar, -30.0, 20.0)
+        self.deterministic = deterministic
+        self.std = torch.exp(0.5 * self.logvar)
+        self.var = torch.exp(self.logvar)
+        if self.deterministic:
+            self.var = self.std = torch.zeros_like(self.mean, device=self.parameters.device)
+
+    def sample(self):
+        x = self.mean + self.std * torch.randn(self.mean.shape, device=self.parameters.device)
+        return x
+
+    def kl(self, other=None):
+        if self.deterministic:
+            return torch.Tensor([0.])
+        else:
+            if other is None:
+                return 0.5 * torch.sum(torch.pow(self.mean, 2)
+                                       + self.var - 1.0 - self.logvar,
+                                       dim=[1, 2, 3])
+            else:
+                return 0.5 * torch.sum(
+                    torch.pow(self.mean - other.mean, 2) / other.var
+                    + self.var / other.var - 1.0 - self.logvar + other.logvar,
+                    dim=[1, 2, 3])
+
+    def nll(self, sample, dims=[1,2,3]):
+        if self.deterministic:
+            return torch.Tensor([0.])
+        logtwopi = np.log(2.0 * np.pi)
+        return 0.5 * torch.sum(
+            logtwopi + self.logvar + torch.pow(sample - self.mean, 2) / self.var,
+            dim=dims)
+
+    def mode(self):
+        return self.mean
+
+
+def normal_kl(mean1, logvar1, mean2, logvar2):
+    """
+    source: https://github.com/openai/guided-diffusion/blob/27c20a8fab9cb472df5d6bdd6c8d11c8f430b924/guided_diffusion/losses.py#L12
+    Compute the KL divergence between two gaussians.
+    Shapes are automatically broadcasted, so batches can be compared to
+    scalars, among other use cases.
+    """
+    tensor = None
+    for obj in (mean1, logvar1, mean2, logvar2):
+        if isinstance(obj, torch.Tensor):
+            tensor = obj
+            break
+    assert tensor is not None, "at least one argument must be a Tensor"
+
+    # Force variances to be Tensors. Broadcasting helps convert scalars to
+    # Tensors, but it does not work for torch.exp().
+    logvar1, logvar2 = [
+        x if isinstance(x, torch.Tensor) else torch.tensor(x).to(tensor)
+        for x in (logvar1, logvar2)
+    ]
+
+    return 0.5 * (
+        -1.0
+        + logvar2
+        - logvar1
+        + torch.exp(logvar1 - logvar2)
+        + ((mean1 - mean2) ** 2) * torch.exp(-logvar2)
+    )

comfy/ldm/mmaudio/vae/vae.py

@@ -0,0 +1,358 @@
+import logging
+from typing import Optional
+
+import torch
+import torch.nn as nn
+
+from .vae_modules import (AttnBlock1D, Downsample1D, ResnetBlock1D,
+                                                 Upsample1D, nonlinearity)
+from .distributions import DiagonalGaussianDistribution
+
+import comfy.ops
+ops = comfy.ops.disable_weight_init
+
+log = logging.getLogger()
+
+DATA_MEAN_80D = [
+    -1.6058, -1.3676, -1.2520, -1.2453, -1.2078, -1.2224, -1.2419, -1.2439, -1.2922, -1.2927,
+    -1.3170, -1.3543, -1.3401, -1.3836, -1.3907, -1.3912, -1.4313, -1.4152, -1.4527, -1.4728,
+    -1.4568, -1.5101, -1.5051, -1.5172, -1.5623, -1.5373, -1.5746, -1.5687, -1.6032, -1.6131,
+    -1.6081, -1.6331, -1.6489, -1.6489, -1.6700, -1.6738, -1.6953, -1.6969, -1.7048, -1.7280,
+    -1.7361, -1.7495, -1.7658, -1.7814, -1.7889, -1.8064, -1.8221, -1.8377, -1.8417, -1.8643,
+    -1.8857, -1.8929, -1.9173, -1.9379, -1.9531, -1.9673, -1.9824, -2.0042, -2.0215, -2.0436,
+    -2.0766, -2.1064, -2.1418, -2.1855, -2.2319, -2.2767, -2.3161, -2.3572, -2.3954, -2.4282,
+    -2.4659, -2.5072, -2.5552, -2.6074, -2.6584, -2.7107, -2.7634, -2.8266, -2.8981, -2.9673
+]
+
+DATA_STD_80D = [
+    1.0291, 1.0411, 1.0043, 0.9820, 0.9677, 0.9543, 0.9450, 0.9392, 0.9343, 0.9297, 0.9276, 0.9263,
+    0.9242, 0.9254, 0.9232, 0.9281, 0.9263, 0.9315, 0.9274, 0.9247, 0.9277, 0.9199, 0.9188, 0.9194,
+    0.9160, 0.9161, 0.9146, 0.9161, 0.9100, 0.9095, 0.9145, 0.9076, 0.9066, 0.9095, 0.9032, 0.9043,
+    0.9038, 0.9011, 0.9019, 0.9010, 0.8984, 0.8983, 0.8986, 0.8961, 0.8962, 0.8978, 0.8962, 0.8973,
+    0.8993, 0.8976, 0.8995, 0.9016, 0.8982, 0.8972, 0.8974, 0.8949, 0.8940, 0.8947, 0.8936, 0.8939,
+    0.8951, 0.8956, 0.9017, 0.9167, 0.9436, 0.9690, 1.0003, 1.0225, 1.0381, 1.0491, 1.0545, 1.0604,
+    1.0761, 1.0929, 1.1089, 1.1196, 1.1176, 1.1156, 1.1117, 1.1070
+]
+
+DATA_MEAN_128D = [
+    -3.3462, -2.6723, -2.4893, -2.3143, -2.2664, -2.3317, -2.1802, -2.4006, -2.2357, -2.4597,
+    -2.3717, -2.4690, -2.5142, -2.4919, -2.6610, -2.5047, -2.7483, -2.5926, -2.7462, -2.7033,
+    -2.7386, -2.8112, -2.7502, -2.9594, -2.7473, -3.0035, -2.8891, -2.9922, -2.9856, -3.0157,
+    -3.1191, -2.9893, -3.1718, -3.0745, -3.1879, -3.2310, -3.1424, -3.2296, -3.2791, -3.2782,
+    -3.2756, -3.3134, -3.3509, -3.3750, -3.3951, -3.3698, -3.4505, -3.4509, -3.5089, -3.4647,
+    -3.5536, -3.5788, -3.5867, -3.6036, -3.6400, -3.6747, -3.7072, -3.7279, -3.7283, -3.7795,
+    -3.8259, -3.8447, -3.8663, -3.9182, -3.9605, -3.9861, -4.0105, -4.0373, -4.0762, -4.1121,
+    -4.1488, -4.1874, -4.2461, -4.3170, -4.3639, -4.4452, -4.5282, -4.6297, -4.7019, -4.7960,
+    -4.8700, -4.9507, -5.0303, -5.0866, -5.1634, -5.2342, -5.3242, -5.4053, -5.4927, -5.5712,
+    -5.6464, -5.7052, -5.7619, -5.8410, -5.9188, -6.0103, -6.0955, -6.1673, -6.2362, -6.3120,
+    -6.3926, -6.4797, -6.5565, -6.6511, -6.8130, -6.9961, -7.1275, -7.2457, -7.3576, -7.4663,
+    -7.6136, -7.7469, -7.8815, -8.0132, -8.1515, -8.3071, -8.4722, -8.7418, -9.3975, -9.6628,
+    -9.7671, -9.8863, -9.9992, -10.0860, -10.1709, -10.5418, -11.2795, -11.3861
+]
+
+DATA_STD_128D = [
+    2.3804, 2.4368, 2.3772, 2.3145, 2.2803, 2.2510, 2.2316, 2.2083, 2.1996, 2.1835, 2.1769, 2.1659,
+    2.1631, 2.1618, 2.1540, 2.1606, 2.1571, 2.1567, 2.1612, 2.1579, 2.1679, 2.1683, 2.1634, 2.1557,
+    2.1668, 2.1518, 2.1415, 2.1449, 2.1406, 2.1350, 2.1313, 2.1415, 2.1281, 2.1352, 2.1219, 2.1182,
+    2.1327, 2.1195, 2.1137, 2.1080, 2.1179, 2.1036, 2.1087, 2.1036, 2.1015, 2.1068, 2.0975, 2.0991,
+    2.0902, 2.1015, 2.0857, 2.0920, 2.0893, 2.0897, 2.0910, 2.0881, 2.0925, 2.0873, 2.0960, 2.0900,
+    2.0957, 2.0958, 2.0978, 2.0936, 2.0886, 2.0905, 2.0845, 2.0855, 2.0796, 2.0840, 2.0813, 2.0817,
+    2.0838, 2.0840, 2.0917, 2.1061, 2.1431, 2.1976, 2.2482, 2.3055, 2.3700, 2.4088, 2.4372, 2.4609,
+    2.4731, 2.4847, 2.5072, 2.5451, 2.5772, 2.6147, 2.6529, 2.6596, 2.6645, 2.6726, 2.6803, 2.6812,
+    2.6899, 2.6916, 2.6931, 2.6998, 2.7062, 2.7262, 2.7222, 2.7158, 2.7041, 2.7485, 2.7491, 2.7451,
+    2.7485, 2.7233, 2.7297, 2.7233, 2.7145, 2.6958, 2.6788, 2.6439, 2.6007, 2.4786, 2.2469, 2.1877,
+    2.1392, 2.0717, 2.0107, 1.9676, 1.9140, 1.7102, 0.9101, 0.7164
+]
+
+
+class VAE(nn.Module):
+
+    def __init__(
+        self,
+        *,
+        data_dim: int,
+        embed_dim: int,
+        hidden_dim: int,
+    ):
+        super().__init__()
+
+        if data_dim == 80:
+            self.data_mean = nn.Buffer(torch.tensor(DATA_MEAN_80D, dtype=torch.float32))
+            self.data_std = nn.Buffer(torch.tensor(DATA_STD_80D, dtype=torch.float32))
+        elif data_dim == 128:
+            self.data_mean = nn.Buffer(torch.tensor(DATA_MEAN_128D, dtype=torch.float32))
+            self.data_std = nn.Buffer(torch.tensor(DATA_STD_128D, dtype=torch.float32))
+
+        self.data_mean = self.data_mean.view(1, -1, 1)
+        self.data_std = self.data_std.view(1, -1, 1)
+
+        self.encoder = Encoder1D(
+            dim=hidden_dim,
+            ch_mult=(1, 2, 4),
+            num_res_blocks=2,
+            attn_layers=[3],
+            down_layers=[0],
+            in_dim=data_dim,
+            embed_dim=embed_dim,
+        )
+        self.decoder = Decoder1D(
+            dim=hidden_dim,
+            ch_mult=(1, 2, 4),
+            num_res_blocks=2,
+            attn_layers=[3],
+            down_layers=[0],
+            in_dim=data_dim,
+            out_dim=data_dim,
+            embed_dim=embed_dim,
+        )
+
+        self.embed_dim = embed_dim
+        # self.quant_conv = nn.Conv1d(2 * embed_dim, 2 * embed_dim, 1)
+        # self.post_quant_conv = nn.Conv1d(embed_dim, embed_dim, 1)
+
+        self.initialize_weights()
+
+    def initialize_weights(self):
+        pass
+
+    def encode(self, x: torch.Tensor, normalize: bool = True) -> DiagonalGaussianDistribution:
+        if normalize:
+            x = self.normalize(x)
+        moments = self.encoder(x)
+        posterior = DiagonalGaussianDistribution(moments)
+        return posterior
+
+    def decode(self, z: torch.Tensor, unnormalize: bool = True) -> torch.Tensor:
+        dec = self.decoder(z)
+        if unnormalize:
+            dec = self.unnormalize(dec)
+        return dec
+
+    def normalize(self, x: torch.Tensor) -> torch.Tensor:
+        return (x - comfy.model_management.cast_to(self.data_mean, dtype=x.dtype, device=x.device)) / comfy.model_management.cast_to(self.data_std, dtype=x.dtype, device=x.device)
+
+    def unnormalize(self, x: torch.Tensor) -> torch.Tensor:
+        return x * comfy.model_management.cast_to(self.data_std, dtype=x.dtype, device=x.device) + comfy.model_management.cast_to(self.data_mean, dtype=x.dtype, device=x.device)
+
+    def forward(
+        self,
+        x: torch.Tensor,
+        sample_posterior: bool = True,
+        rng: Optional[torch.Generator] = None,
+        normalize: bool = True,
+        unnormalize: bool = True,
+    ) -> tuple[torch.Tensor, DiagonalGaussianDistribution]:
+
+        posterior = self.encode(x, normalize=normalize)
+        if sample_posterior:
+            z = posterior.sample(rng)
+        else:
+            z = posterior.mode()
+        dec = self.decode(z, unnormalize=unnormalize)
+        return dec, posterior
+
+    def load_weights(self, src_dict) -> None:
+        self.load_state_dict(src_dict, strict=True)
+
+    @property
+    def device(self) -> torch.device:
+        return next(self.parameters()).device
+
+    def get_last_layer(self):
+        return self.decoder.conv_out.weight
+
+    def remove_weight_norm(self):
+        return self
+
+
+class Encoder1D(nn.Module):
+
+    def __init__(self,
+                 *,
+                 dim: int,
+                 ch_mult: tuple[int] = (1, 2, 4, 8),
+                 num_res_blocks: int,
+                 attn_layers: list[int] = [],
+                 down_layers: list[int] = [],
+                 resamp_with_conv: bool = True,
+                 in_dim: int,
+                 embed_dim: int,
+                 double_z: bool = True,
+                 kernel_size: int = 3,
+                 clip_act: float = 256.0):
+        super().__init__()
+        self.dim = dim
+        self.num_layers = len(ch_mult)
+        self.num_res_blocks = num_res_blocks
+        self.in_channels = in_dim
+        self.clip_act = clip_act
+        self.down_layers = down_layers
+        self.attn_layers = attn_layers
+        self.conv_in = ops.Conv1d(in_dim, self.dim, kernel_size=kernel_size, padding=kernel_size // 2, bias=False)
+
+        in_ch_mult = (1, ) + tuple(ch_mult)
+        self.in_ch_mult = in_ch_mult
+        # downsampling
+        self.down = nn.ModuleList()
+        for i_level in range(self.num_layers):
+            block = nn.ModuleList()
+            attn = nn.ModuleList()
+            block_in = dim * in_ch_mult[i_level]
+            block_out = dim * ch_mult[i_level]
+            for i_block in range(self.num_res_blocks):
+                block.append(
+                    ResnetBlock1D(in_dim=block_in,
+                                  out_dim=block_out,
+                                  kernel_size=kernel_size,
+                                  use_norm=True))
+                block_in = block_out
+                if i_level in attn_layers:
+                    attn.append(AttnBlock1D(block_in))
+            down = nn.Module()
+            down.block = block
+            down.attn = attn
+            if i_level in down_layers:
+                down.downsample = Downsample1D(block_in, resamp_with_conv)
+            self.down.append(down)
+
+        # middle
+        self.mid = nn.Module()
+        self.mid.block_1 = ResnetBlock1D(in_dim=block_in,
+                                         out_dim=block_in,
+                                         kernel_size=kernel_size,
+                                         use_norm=True)
+        self.mid.attn_1 = AttnBlock1D(block_in)
+        self.mid.block_2 = ResnetBlock1D(in_dim=block_in,
+                                         out_dim=block_in,
+                                         kernel_size=kernel_size,
+                                         use_norm=True)
+
+        # end
+        self.conv_out = ops.Conv1d(block_in,
+                                 2 * embed_dim if double_z else embed_dim,
+                                 kernel_size=kernel_size, padding=kernel_size // 2, bias=False)
+
+        self.learnable_gain = nn.Parameter(torch.zeros([]))
+
+    def forward(self, x):
+
+        # downsampling
+        h = self.conv_in(x)
+        for i_level in range(self.num_layers):
+            for i_block in range(self.num_res_blocks):
+                h = self.down[i_level].block[i_block](h)
+                if len(self.down[i_level].attn) > 0:
+                    h = self.down[i_level].attn[i_block](h)
+                h = h.clamp(-self.clip_act, self.clip_act)
+            if i_level in self.down_layers:
+                h = self.down[i_level].downsample(h)
+
+        # middle
+        h = self.mid.block_1(h)
+        h = self.mid.attn_1(h)
+        h = self.mid.block_2(h)
+        h = h.clamp(-self.clip_act, self.clip_act)
+
+        # end
+        h = nonlinearity(h)
+        h = self.conv_out(h) * (self.learnable_gain + 1)
+        return h
+
+
+class Decoder1D(nn.Module):
+
+    def __init__(self,
+                 *,
+                 dim: int,
+                 out_dim: int,
+                 ch_mult: tuple[int] = (1, 2, 4, 8),
+                 num_res_blocks: int,
+                 attn_layers: list[int] = [],
+                 down_layers: list[int] = [],
+                 kernel_size: int = 3,
+                 resamp_with_conv: bool = True,
+                 in_dim: int,
+                 embed_dim: int,
+                 clip_act: float = 256.0):
+        super().__init__()
+        self.ch = dim
+        self.num_layers = len(ch_mult)
+        self.num_res_blocks = num_res_blocks
+        self.in_channels = in_dim
+        self.clip_act = clip_act
+        self.down_layers = [i + 1 for i in down_layers]  # each downlayer add one
+
+        # compute in_ch_mult, block_in and curr_res at lowest res
+        block_in = dim * ch_mult[self.num_layers - 1]
+
+        # z to block_in
+        self.conv_in = ops.Conv1d(embed_dim, block_in, kernel_size=kernel_size, padding=kernel_size // 2, bias=False)
+
+        # middle
+        self.mid = nn.Module()
+        self.mid.block_1 = ResnetBlock1D(in_dim=block_in, out_dim=block_in, use_norm=True)
+        self.mid.attn_1 = AttnBlock1D(block_in)
+        self.mid.block_2 = ResnetBlock1D(in_dim=block_in, out_dim=block_in, use_norm=True)
+
+        # upsampling
+        self.up = nn.ModuleList()
+        for i_level in reversed(range(self.num_layers)):
+            block = nn.ModuleList()
+            attn = nn.ModuleList()
+            block_out = dim * ch_mult[i_level]
+            for i_block in range(self.num_res_blocks + 1):
+                block.append(ResnetBlock1D(in_dim=block_in, out_dim=block_out, use_norm=True))
+                block_in = block_out
+                if i_level in attn_layers:
+                    attn.append(AttnBlock1D(block_in))
+            up = nn.Module()
+            up.block = block
+            up.attn = attn
+            if i_level in self.down_layers:
+                up.upsample = Upsample1D(block_in, resamp_with_conv)
+            self.up.insert(0, up)  # prepend to get consistent order
+
+        # end
+        self.conv_out = ops.Conv1d(block_in, out_dim, kernel_size=kernel_size, padding=kernel_size // 2, bias=False)
+        self.learnable_gain = nn.Parameter(torch.zeros([]))
+
+    def forward(self, z):
+        # z to block_in
+        h = self.conv_in(z)
+
+        # middle
+        h = self.mid.block_1(h)
+        h = self.mid.attn_1(h)
+        h = self.mid.block_2(h)
+        h = h.clamp(-self.clip_act, self.clip_act)
+
+        # upsampling
+        for i_level in reversed(range(self.num_layers)):
+            for i_block in range(self.num_res_blocks + 1):
+                h = self.up[i_level].block[i_block](h)
+                if len(self.up[i_level].attn) > 0:
+                    h = self.up[i_level].attn[i_block](h)
+                h = h.clamp(-self.clip_act, self.clip_act)
+            if i_level in self.down_layers:
+                h = self.up[i_level].upsample(h)
+
+        h = nonlinearity(h)
+        h = self.conv_out(h) * (self.learnable_gain + 1)
+        return h
+
+
+def VAE_16k(**kwargs) -> VAE:
+    return VAE(data_dim=80, embed_dim=20, hidden_dim=384, **kwargs)
+
+
+def VAE_44k(**kwargs) -> VAE:
+    return VAE(data_dim=128, embed_dim=40, hidden_dim=512, **kwargs)
+
+
+def get_my_vae(name: str, **kwargs) -> VAE:
+    if name == '16k':
+        return VAE_16k(**kwargs)
+    if name == '44k':
+        return VAE_44k(**kwargs)
+    raise ValueError(f'Unknown model: {name}')
+

comfy/ldm/mmaudio/vae/vae_modules.py

@@ -0,0 +1,121 @@
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from comfy.ldm.modules.diffusionmodules.model import vae_attention
+import math
+import comfy.ops
+ops = comfy.ops.disable_weight_init
+
+def nonlinearity(x):
+    # swish
+    return torch.nn.functional.silu(x) / 0.596
+
+def mp_sum(a, b, t=0.5):
+    return a.lerp(b, t) / math.sqrt((1 - t)**2 + t**2)
+
+def normalize(x, dim=None, eps=1e-4):
+    if dim is None:
+        dim = list(range(1, x.ndim))
+    norm = torch.linalg.vector_norm(x, dim=dim, keepdim=True, dtype=torch.float32)
+    norm = torch.add(eps, norm, alpha=math.sqrt(norm.numel() / x.numel()))
+    return x / norm.to(x.dtype)
+
+class ResnetBlock1D(nn.Module):
+
+    def __init__(self, *, in_dim, out_dim=None, conv_shortcut=False, kernel_size=3, use_norm=True):
+        super().__init__()
+        self.in_dim = in_dim
+        out_dim = in_dim if out_dim is None else out_dim
+        self.out_dim = out_dim
+        self.use_conv_shortcut = conv_shortcut
+        self.use_norm = use_norm
+
+        self.conv1 = ops.Conv1d(in_dim, out_dim, kernel_size=kernel_size, padding=kernel_size // 2, bias=False)
+        self.conv2 = ops.Conv1d(out_dim, out_dim, kernel_size=kernel_size, padding=kernel_size // 2, bias=False)
+        if self.in_dim != self.out_dim:
+            if self.use_conv_shortcut:
+                self.conv_shortcut = ops.Conv1d(in_dim, out_dim, kernel_size=kernel_size, padding=kernel_size // 2, bias=False)
+            else:
+                self.nin_shortcut = ops.Conv1d(in_dim, out_dim, kernel_size=1, padding=0, bias=False)
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+
+        # pixel norm
+        if self.use_norm:
+            x = normalize(x, dim=1)
+
+        h = x
+        h = nonlinearity(h)
+        h = self.conv1(h)
+
+        h = nonlinearity(h)
+        h = self.conv2(h)
+
+        if self.in_dim != self.out_dim:
+            if self.use_conv_shortcut:
+                x = self.conv_shortcut(x)
+            else:
+                x = self.nin_shortcut(x)
+
+        return mp_sum(x, h, t=0.3)
+
+
+class AttnBlock1D(nn.Module):
+
+    def __init__(self, in_channels, num_heads=1):
+        super().__init__()
+        self.in_channels = in_channels
+
+        self.num_heads = num_heads
+        self.qkv = ops.Conv1d(in_channels, in_channels * 3, kernel_size=1, padding=0, bias=False)
+        self.proj_out = ops.Conv1d(in_channels, in_channels, kernel_size=1, padding=0, bias=False)
+        self.optimized_attention = vae_attention()
+
+    def forward(self, x):
+        h = x
+        y = self.qkv(h)
+        y = y.reshape(y.shape[0], -1, 3, y.shape[-1])
+        q, k, v = normalize(y, dim=1).unbind(2)
+
+        h = self.optimized_attention(q, k, v)
+        h = self.proj_out(h)
+
+        return mp_sum(x, h, t=0.3)
+
+
+class Upsample1D(nn.Module):
+
+    def __init__(self, in_channels, with_conv):
+        super().__init__()
+        self.with_conv = with_conv
+        if self.with_conv:
+            self.conv = ops.Conv1d(in_channels, in_channels, kernel_size=3, padding=1, bias=False)
+
+    def forward(self, x):
+        x = F.interpolate(x, scale_factor=2.0, mode='nearest-exact')  # support 3D tensor(B,C,T)
+        if self.with_conv:
+            x = self.conv(x)
+        return x
+
+
+class Downsample1D(nn.Module):
+
+    def __init__(self, in_channels, with_conv):
+        super().__init__()
+        self.with_conv = with_conv
+        if self.with_conv:
+            # no asymmetric padding in torch conv, must do it ourselves
+            self.conv1 = ops.Conv1d(in_channels, in_channels, kernel_size=1, padding=0, bias=False)
+            self.conv2 = ops.Conv1d(in_channels, in_channels, kernel_size=1, padding=0, bias=False)
+
+    def forward(self, x):
+
+        if self.with_conv:
+            x = self.conv1(x)
+
+        x = F.avg_pool1d(x, kernel_size=2, stride=2)
+
+        if self.with_conv:
+            x = self.conv2(x)
+
+        return x

comfy/ldm/wan/vae2_2.py

@@ -657,51 +657,51 @@ class WanVAE(nn.Module):
         )
 
     def encode(self, x):
-        self.clear_cache()
+        conv_idx = [0]
+        feat_map = [None] * count_conv3d(self.encoder)
         x = patchify(x, patch_size=2)
         t = x.shape[2]
         iter_ = 1 + (t - 1) // 4
         for i in range(iter_):
-            self._enc_conv_idx = [0]
+            conv_idx = [0]
             if i == 0:
                 out = self.encoder(
                     x[:, :, :1, :, :],
-                    feat_cache=self._enc_feat_map,
-                    feat_idx=self._enc_conv_idx,
+                    feat_cache=feat_map,
+                    feat_idx=conv_idx,
                 )
             else:
                 out_ = self.encoder(
                     x[:, :, 1 + 4 * (i - 1):1 + 4 * i, :, :],
-                    feat_cache=self._enc_feat_map,
-                    feat_idx=self._enc_conv_idx,
+                    feat_cache=feat_map,
+                    feat_idx=conv_idx,
                 )
                 out = torch.cat([out, out_], 2)
         mu, log_var = self.conv1(out).chunk(2, dim=1)
-        self.clear_cache()
         return mu
 
     def decode(self, z):
-        self.clear_cache()
+        conv_idx = [0]
+        feat_map = [None] * count_conv3d(self.decoder)
         iter_ = z.shape[2]
         x = self.conv2(z)
         for i in range(iter_):
-            self._conv_idx = [0]
+            conv_idx = [0]
             if i == 0:
                 out = self.decoder(
                     x[:, :, i:i + 1, :, :],
-                    feat_cache=self._feat_map,
-                    feat_idx=self._conv_idx,
+                    feat_cache=feat_map,
+                    feat_idx=conv_idx,
                     first_chunk=True,
                 )
             else:
                 out_ = self.decoder(
                     x[:, :, i:i + 1, :, :],
-                    feat_cache=self._feat_map,
-                    feat_idx=self._conv_idx,
+                    feat_cache=feat_map,
+                    feat_idx=conv_idx,
                 )
                 out = torch.cat([out, out_], 2)
         out = unpatchify(out, patch_size=2)
-        self.clear_cache()
         return out
 
     def reparameterize(self, mu, log_var):
@@ -715,12 +715,3 @@ class WanVAE(nn.Module):
             return mu
         std = torch.exp(0.5 * log_var.clamp(-30.0, 20.0))
         return mu + std * torch.randn_like(std)
-
-    def clear_cache(self):
-        self._conv_num = count_conv3d(self.decoder)
-        self._conv_idx = [0]
-        self._feat_map = [None] * self._conv_num
-        # cache encode
-        self._enc_conv_num = count_conv3d(self.encoder)
-        self._enc_conv_idx = [0]
-        self._enc_feat_map = [None] * self._enc_conv_num

comfy/model_base.py

@@ -138,6 +138,7 @@ class BaseModel(torch.nn.Module):
             else:
                 operations = model_config.custom_operations
             self.diffusion_model = unet_model(**unet_config, device=device, operations=operations)
+            self.diffusion_model.eval()
             if comfy.model_management.force_channels_last():
                 self.diffusion_model.to(memory_format=torch.channels_last)
                 logging.debug("using channels last mode for diffusion model")
@@ -669,7 +670,6 @@ class Lotus(BaseModel):
 class StableCascade_C(BaseModel):
     def __init__(self, model_config, model_type=ModelType.STABLE_CASCADE, device=None):
         super().__init__(model_config, model_type, device=device, unet_model=StageC)
-        self.diffusion_model.eval().requires_grad_(False)
 
     def extra_conds(self, **kwargs):
         out = {}
@@ -698,7 +698,6 @@ class StableCascade_C(BaseModel):
 class StableCascade_B(BaseModel):
     def __init__(self, model_config, model_type=ModelType.STABLE_CASCADE, device=None):
         super().__init__(model_config, model_type, device=device, unet_model=StageB)
-        self.diffusion_model.eval().requires_grad_(False)
 
     def extra_conds(self, **kwargs):
         out = {}

comfy/model_detection.py

@@ -213,7 +213,7 @@ def detect_unet_config(state_dict, key_prefix, metadata=None):
                 dit_config["nerf_mlp_ratio"] = 4
                 dit_config["nerf_depth"] = 4
                 dit_config["nerf_max_freqs"] = 8
-                dit_config["nerf_tile_size"] = 32
+                dit_config["nerf_tile_size"] = 512
                 dit_config["nerf_final_head_type"] = "conv" if f"{key_prefix}nerf_final_layer_conv.norm.scale" in state_dict_keys else "linear"
                 dit_config["nerf_embedder_dtype"] = torch.float32
         else:
@@ -365,8 +365,8 @@ def detect_unet_config(state_dict, key_prefix, metadata=None):
         dit_config["patch_size"] = 2
         dit_config["in_channels"] = 16
         dit_config["dim"] = 2304
-        dit_config["cap_feat_dim"] = 2304
-        dit_config["n_layers"] = 26
+        dit_config["cap_feat_dim"] = state_dict['{}cap_embedder.1.weight'.format(key_prefix)].shape[1]
+        dit_config["n_layers"] = count_blocks(state_dict_keys, '{}layers.'.format(key_prefix) + '{}.')
         dit_config["n_heads"] = 24
         dit_config["n_kv_heads"] = 8
         dit_config["qk_norm"] = True

comfy/model_management.py

@@ -332,6 +332,8 @@ except:
 SUPPORT_FP8_OPS = args.supports_fp8_compute
 try:
     if is_amd():
+        torch.backends.cudnn.enabled = False  # Seems to improve things a lot on AMD
+        logging.info("Set: torch.backends.cudnn.enabled = False for better AMD performance.")
         try:
             rocm_version = tuple(map(int, str(torch.version.hip).split(".")[:2]))
         except:
@@ -344,11 +346,11 @@ try:
                 if torch_version_numeric >= (2, 7):  # works on 2.6 but doesn't actually seem to improve much
                     if any((a in arch) for a in ["gfx90a", "gfx942", "gfx1100", "gfx1101", "gfx1151"]):  # TODO: more arches, TODO: gfx950
                         ENABLE_PYTORCH_ATTENTION = True
-#                if torch_version_numeric >= (2, 8):
-#                    if any((a in arch) for a in ["gfx1201"]):
-#                        ENABLE_PYTORCH_ATTENTION = True
+                if rocm_version >= (7, 0):
+                   if any((a in arch) for a in ["gfx1201"]):
+                       ENABLE_PYTORCH_ATTENTION = True
         if torch_version_numeric >= (2, 7) and rocm_version >= (6, 4):
-            if any((a in arch) for a in ["gfx1200", "gfx1201", "gfx942", "gfx950"]):  # TODO: more arches
+            if any((a in arch) for a in ["gfx1200", "gfx1201", "gfx950"]):  # TODO: more arches, "gfx942" gives error on pytorch nightly 2.10 1013 rocm7.0
                 SUPPORT_FP8_OPS = True
 
 except:
@@ -370,6 +372,9 @@ try:
 except:
     pass
 
+if torch.cuda.is_available() and torch.backends.cudnn.is_available() and PerformanceFeature.AutoTune in args.fast:
+    torch.backends.cudnn.benchmark = True
+
 try:
     if torch_version_numeric >= (2, 5):
         torch.backends.cuda.allow_fp16_bf16_reduction_math_sdp(True)
@@ -925,11 +930,7 @@ def vae_dtype(device=None, allowed_dtypes=[]):
         if d == torch.float16 and should_use_fp16(device):
             return d
 
-        # NOTE: bfloat16 seems to work on AMD for the VAE but is extremely slow in some cases compared to fp32
-        # slowness still a problem on pytorch nightly 2.9.0.dev20250720+rocm6.4 tested on RDNA3
-        # also a problem on RDNA4 except fp32 is also slow there.
-        # This is due to large bf16 convolutions being extremely slow.
-        if d == torch.bfloat16 and ((not is_amd()) or amd_min_version(device, min_rdna_version=4)) and should_use_bf16(device):
+        if d == torch.bfloat16 and should_use_bf16(device):
             return d
 
     return torch.float32

comfy/model_patcher.py

@@ -123,16 +123,30 @@ def move_weight_functions(m, device):
     return memory
 
 class LowVramPatch:
-    def __init__(self, key, patches):
+    def __init__(self, key, patches, convert_func=None, set_func=None):
         self.key = key
         self.patches = patches
+        self.convert_func = convert_func
+        self.set_func = set_func
+
     def __call__(self, weight):
         intermediate_dtype = weight.dtype
+        if self.convert_func is not None:
+            weight = self.convert_func(weight.to(dtype=torch.float32, copy=True), inplace=True)
+
         if intermediate_dtype not in [torch.float32, torch.float16, torch.bfloat16]: #intermediate_dtype has to be one that is supported in math ops
             intermediate_dtype = torch.float32
-            return comfy.float.stochastic_rounding(comfy.lora.calculate_weight(self.patches[self.key], weight.to(intermediate_dtype), self.key, intermediate_dtype=intermediate_dtype), weight.dtype, seed=string_to_seed(self.key))
+            out = comfy.lora.calculate_weight(self.patches[self.key], weight.to(intermediate_dtype), self.key, intermediate_dtype=intermediate_dtype)
+            if self.set_func is None:
+                return comfy.float.stochastic_rounding(out, weight.dtype, seed=string_to_seed(self.key))
+            else:
+                return self.set_func(out, seed=string_to_seed(self.key), return_weight=True)
 
-        return comfy.lora.calculate_weight(self.patches[self.key], weight, self.key, intermediate_dtype=intermediate_dtype)
+        out = comfy.lora.calculate_weight(self.patches[self.key], weight, self.key, intermediate_dtype=intermediate_dtype)
+        if self.set_func is not None:
+            return self.set_func(out, seed=string_to_seed(self.key), return_weight=True).to(dtype=intermediate_dtype)
+        else:
+            return out
 
 def get_key_weight(model, key):
     set_func = None
@@ -657,13 +671,15 @@ class ModelPatcher:
                         if force_patch_weights:
                             self.patch_weight_to_device(weight_key)
                         else:
-                            m.weight_function = [LowVramPatch(weight_key, self.patches)]
+                            _, set_func, convert_func = get_key_weight(self.model, weight_key)
+                            m.weight_function = [LowVramPatch(weight_key, self.patches, convert_func, set_func)]
                             patch_counter += 1
                     if bias_key in self.patches:
                         if force_patch_weights:
                             self.patch_weight_to_device(bias_key)
                         else:
-                            m.bias_function = [LowVramPatch(bias_key, self.patches)]
+                            _, set_func, convert_func = get_key_weight(self.model, bias_key)
+                            m.bias_function = [LowVramPatch(bias_key, self.patches, convert_func, set_func)]
                             patch_counter += 1
 
                     cast_weight = True
@@ -825,10 +841,12 @@ class ModelPatcher:
                         module_mem += move_weight_functions(m, device_to)
                         if lowvram_possible:
                             if weight_key in self.patches:
-                                m.weight_function.append(LowVramPatch(weight_key, self.patches))
+                                _, set_func, convert_func = get_key_weight(self.model, weight_key)
+                                m.weight_function.append(LowVramPatch(weight_key, self.patches, convert_func, set_func))
                                 patch_counter += 1
                             if bias_key in self.patches:
-                                m.bias_function.append(LowVramPatch(bias_key, self.patches))
+                                _, set_func, convert_func = get_key_weight(self.model, bias_key)
+                                m.bias_function.append(LowVramPatch(bias_key, self.patches, convert_func, set_func))
                                 patch_counter += 1
                             cast_weight = True
 

comfy/model_sampling.py

@@ -21,17 +21,23 @@ def rescale_zero_terminal_snr_sigmas(sigmas):
     alphas_bar[-1] = 4.8973451890853435e-08
     return ((1 - alphas_bar) / alphas_bar) ** 0.5
 
+def reshape_sigma(sigma, noise_dim):
+    if sigma.nelement() == 1:
+        return sigma.view(())
+    else:
+        return sigma.view(sigma.shape[:1] + (1,) * (noise_dim - 1))
+
 class EPS:
     def calculate_input(self, sigma, noise):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (noise.ndim - 1))
+        sigma = reshape_sigma(sigma, noise.ndim)
         return noise / (sigma ** 2 + self.sigma_data ** 2) ** 0.5
 
     def calculate_denoised(self, sigma, model_output, model_input):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (model_output.ndim - 1))
+        sigma = reshape_sigma(sigma, model_output.ndim)
         return model_input - model_output * sigma
 
     def noise_scaling(self, sigma, noise, latent_image, max_denoise=False):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (noise.ndim - 1))
+        sigma = reshape_sigma(sigma, noise.ndim)
         if max_denoise:
             noise = noise * torch.sqrt(1.0 + sigma ** 2.0)
         else:
@@ -45,12 +51,12 @@ class EPS:
 
 class V_PREDICTION(EPS):
     def calculate_denoised(self, sigma, model_output, model_input):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (model_output.ndim - 1))
+        sigma = reshape_sigma(sigma, model_output.ndim)
         return model_input * self.sigma_data ** 2 / (sigma ** 2 + self.sigma_data ** 2) - model_output * sigma * self.sigma_data / (sigma ** 2 + self.sigma_data ** 2) ** 0.5
 
 class EDM(V_PREDICTION):
     def calculate_denoised(self, sigma, model_output, model_input):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (model_output.ndim - 1))
+        sigma = reshape_sigma(sigma, model_output.ndim)
         return model_input * self.sigma_data ** 2 / (sigma ** 2 + self.sigma_data ** 2) + model_output * sigma * self.sigma_data / (sigma ** 2 + self.sigma_data ** 2) ** 0.5
 
 class CONST:
@@ -58,15 +64,15 @@ class CONST:
         return noise
 
     def calculate_denoised(self, sigma, model_output, model_input):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (model_output.ndim - 1))
+        sigma = reshape_sigma(sigma, model_output.ndim)
         return model_input - model_output * sigma
 
     def noise_scaling(self, sigma, noise, latent_image, max_denoise=False):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (noise.ndim - 1))
+        sigma = reshape_sigma(sigma, noise.ndim)
         return sigma * noise + (1.0 - sigma) * latent_image
 
     def inverse_noise_scaling(self, sigma, latent):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (latent.ndim - 1))
+        sigma = reshape_sigma(sigma, latent.ndim)
         return latent / (1.0 - sigma)
 
 class X0(EPS):
@@ -80,16 +86,16 @@ class IMG_TO_IMG(X0):
 class COSMOS_RFLOW:
     def calculate_input(self, sigma, noise):
         sigma = (sigma / (sigma + 1))
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (noise.ndim - 1))
+        sigma = reshape_sigma(sigma, noise.ndim)
         return noise * (1.0 - sigma)
 
     def calculate_denoised(self, sigma, model_output, model_input):
         sigma = (sigma / (sigma + 1))
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (model_output.ndim - 1))
+        sigma = reshape_sigma(sigma, model_output.ndim)
         return model_input * (1.0 - sigma) - model_output * sigma
 
     def noise_scaling(self, sigma, noise, latent_image, max_denoise=False):
-        sigma = sigma.view(sigma.shape[:1] + (1,) * (noise.ndim - 1))
+        sigma = reshape_sigma(sigma, noise.ndim)
         noise = noise * sigma
         noise += latent_image
         return noise

comfy/ops.py

@@ -24,6 +24,11 @@ import comfy.float
 import comfy.rmsnorm
 import contextlib
 
+def run_every_op():
+    if torch.compiler.is_compiling():
+        return
+
+    comfy.model_management.throw_exception_if_processing_interrupted()
 
 def scaled_dot_product_attention(q, k, v, *args, **kwargs):
     return torch.nn.functional.scaled_dot_product_attention(q, k, v, *args, **kwargs)
@@ -50,14 +55,22 @@ try:
 except (ModuleNotFoundError, TypeError):
     logging.warning("Could not set sdpa backend priority.")
 
-cast_to = comfy.model_management.cast_to #TODO: remove once no more references
+NVIDIA_MEMORY_CONV_BUG_WORKAROUND = False
+try:
+    if comfy.model_management.is_nvidia():
+        if torch.backends.cudnn.version() >= 91002 and comfy.model_management.torch_version_numeric >= (2, 9) and comfy.model_management.torch_version_numeric <= (2, 10):
+            #TODO: change upper bound version once it's fixed'
+            NVIDIA_MEMORY_CONV_BUG_WORKAROUND = True
+            logging.info("working around nvidia conv3d memory bug.")
+except:
+    pass
 
-if torch.cuda.is_available() and torch.backends.cudnn.is_available() and PerformanceFeature.AutoTune in args.fast:
-    torch.backends.cudnn.benchmark = True
+cast_to = comfy.model_management.cast_to #TODO: remove once no more references
 
 def cast_to_input(weight, input, non_blocking=False, copy=True):
     return comfy.model_management.cast_to(weight, input.dtype, input.device, non_blocking=non_blocking, copy=copy)
 
+@torch.compiler.disable()
 def cast_bias_weight(s, input=None, dtype=None, device=None, bias_dtype=None):
     if input is not None:
         if dtype is None:
@@ -109,6 +122,7 @@ class disable_weight_init:
             return torch.nn.functional.linear(input, weight, bias)
 
         def forward(self, *args, **kwargs):
+            run_every_op()
             if self.comfy_cast_weights or len(self.weight_function) > 0 or len(self.bias_function) > 0:
                 return self.forward_comfy_cast_weights(*args, **kwargs)
             else:
@@ -123,6 +137,7 @@ class disable_weight_init:
             return self._conv_forward(input, weight, bias)
 
         def forward(self, *args, **kwargs):
+            run_every_op()
             if self.comfy_cast_weights or len(self.weight_function) > 0 or len(self.bias_function) > 0:
                 return self.forward_comfy_cast_weights(*args, **kwargs)
             else:
@@ -137,6 +152,7 @@ class disable_weight_init:
             return self._conv_forward(input, weight, bias)
 
         def forward(self, *args, **kwargs):
+            run_every_op()
             if self.comfy_cast_weights or len(self.weight_function) > 0 or len(self.bias_function) > 0:
                 return self.forward_comfy_cast_weights(*args, **kwargs)
             else:
@@ -146,11 +162,21 @@ class disable_weight_init:
         def reset_parameters(self):
             return None
 
+        def _conv_forward(self, input, weight, bias, *args, **kwargs):
+            if NVIDIA_MEMORY_CONV_BUG_WORKAROUND and weight.dtype in (torch.float16, torch.bfloat16):
+                out = torch.cudnn_convolution(input, weight, self.padding, self.stride, self.dilation, self.groups, benchmark=False, deterministic=False, allow_tf32=True)
+                if bias is not None:
+                    out += bias.reshape((1, -1) + (1,) * (out.ndim - 2))
+                return out
+            else:
+                return super()._conv_forward(input, weight, bias, *args, **kwargs)
+
         def forward_comfy_cast_weights(self, input):
             weight, bias = cast_bias_weight(self, input)
             return self._conv_forward(input, weight, bias)
 
         def forward(self, *args, **kwargs):
+            run_every_op()
             if self.comfy_cast_weights or len(self.weight_function) > 0 or len(self.bias_function) > 0:
                 return self.forward_comfy_cast_weights(*args, **kwargs)
             else:
@@ -165,6 +191,7 @@ class disable_weight_init:
             return torch.nn.functional.group_norm(input, self.num_groups, weight, bias, self.eps)
 
         def forward(self, *args, **kwargs):
+            run_every_op()
             if self.comfy_cast_weights or len(self.weight_function) > 0 or len(self.bias_function) > 0:
                 return self.forward_comfy_cast_weights(*args, **kwargs)
             else:
@@ -183,6 +210,7 @@ class disable_weight_init:
             return torch.nn.functional.layer_norm(input, self.normalized_shape, weight, bias, self.eps)
 
         def forward(self, *args, **kwargs):
+            run_every_op()
             if self.comfy_cast_weights or len(self.weight_function) > 0 or len(self.bias_function) > 0:
                 return self.forward_comfy_cast_weights(*args, **kwargs)
             else:
@@ -202,6 +230,7 @@ class disable_weight_init:
             # return torch.nn.functional.rms_norm(input, self.normalized_shape, weight, self.eps)
 
         def forward(self, *args, **kwargs):
+            run_every_op()
             if self.comfy_cast_weights or len(self.weight_function) > 0 or len(self.bias_function) > 0:
                 return self.forward_comfy_cast_weights(*args, **kwargs)
             else:
@@ -223,6 +252,7 @@ class disable_weight_init:
                 output_padding, self.groups, self.dilation)
 
         def forward(self, *args, **kwargs):
+            run_every_op()
             if self.comfy_cast_weights or len(self.weight_function) > 0 or len(self.bias_function) > 0:
                 return self.forward_comfy_cast_weights(*args, **kwargs)
             else:
@@ -244,6 +274,7 @@ class disable_weight_init:
                 output_padding, self.groups, self.dilation)
 
         def forward(self, *args, **kwargs):
+            run_every_op()
             if self.comfy_cast_weights or len(self.weight_function) > 0 or len(self.bias_function) > 0:
                 return self.forward_comfy_cast_weights(*args, **kwargs)
             else:
@@ -262,6 +293,7 @@ class disable_weight_init:
             return torch.nn.functional.embedding(input, weight, self.padding_idx, self.max_norm, self.norm_type, self.scale_grad_by_freq, self.sparse).to(dtype=output_dtype)
 
         def forward(self, *args, **kwargs):
+            run_every_op()
             if self.comfy_cast_weights or len(self.weight_function) > 0 or len(self.bias_function) > 0:
                 return self.forward_comfy_cast_weights(*args, **kwargs)
             else:
@@ -416,8 +448,10 @@ def scaled_fp8_ops(fp8_matrix_mult=False, scale_input=False, override_dtype=None
                 else:
                     return weight * self.scale_weight.to(device=weight.device, dtype=weight.dtype)
 
-            def set_weight(self, weight, inplace_update=False, seed=None, **kwargs):
+            def set_weight(self, weight, inplace_update=False, seed=None, return_weight=False, **kwargs):
                 weight = comfy.float.stochastic_rounding(weight / self.scale_weight.to(device=weight.device, dtype=weight.dtype), self.weight.dtype, seed=seed)
+                if return_weight:
+                    return weight
                 if inplace_update:
                     self.weight.data.copy_(weight)
                 else:

comfy/patcher_extension.py

@@ -150,7 +150,7 @@ def merge_nested_dicts(dict1: dict, dict2: dict, copy_dict1=True):
     for key, value in dict2.items():
         if isinstance(value, dict):
             curr_value = merged_dict.setdefault(key, {})
-            merged_dict[key] = merge_nested_dicts(value, curr_value)
+            merged_dict[key] = merge_nested_dicts(curr_value, value)
         elif isinstance(value, list):
             merged_dict.setdefault(key, []).extend(value)
         else:

comfy/samplers.py

@@ -306,17 +306,10 @@ def _calc_cond_batch(model: BaseModel, conds: list[list[dict]], x_in: torch.Tens
                                                                                  copy_dict1=False)
 
             if patches is not None:
-                # TODO: replace with merge_nested_dicts function
-                if "patches" in transformer_options:
-                    cur_patches = transformer_options["patches"].copy()
-                    for p in patches:
-                        if p in cur_patches:
-                            cur_patches[p] = cur_patches[p] + patches[p]
-                        else:
-                            cur_patches[p] = patches[p]
-                    transformer_options["patches"] = cur_patches
-                else:
-                    transformer_options["patches"] = patches
+                transformer_options["patches"] = comfy.patcher_extension.merge_nested_dicts(
+                    transformer_options.get("patches", {}),
+                    patches
+                )
 
             transformer_options["cond_or_uncond"] = cond_or_uncond[:]
             transformer_options["uuids"] = uuids[:]

comfy/sd.py

@@ -18,6 +18,7 @@ import comfy.ldm.wan.vae2_2
 import comfy.ldm.hunyuan3d.vae
 import comfy.ldm.ace.vae.music_dcae_pipeline
 import comfy.ldm.hunyuan_video.vae
+import comfy.ldm.mmaudio.vae.autoencoder
 import comfy.pixel_space_convert
 import yaml
 import math
@@ -275,8 +276,13 @@ class VAE:
         if 'decoder.up_blocks.0.resnets.0.norm1.weight' in sd.keys(): #diffusers format
             sd = diffusers_convert.convert_vae_state_dict(sd)
 
-        self.memory_used_encode = lambda shape, dtype: (1767 * shape[2] * shape[3]) * model_management.dtype_size(dtype) #These are for AutoencoderKL and need tweaking (should be lower)
-        self.memory_used_decode = lambda shape, dtype: (2178 * shape[2] * shape[3] * 64) * model_management.dtype_size(dtype)
+        if model_management.is_amd():
+            VAE_KL_MEM_RATIO = 2.73
+        else:
+            VAE_KL_MEM_RATIO = 1.0
+
+        self.memory_used_encode = lambda shape, dtype: (1767 * shape[2] * shape[3]) * model_management.dtype_size(dtype) * VAE_KL_MEM_RATIO #These are for AutoencoderKL and need tweaking (should be lower)
+        self.memory_used_decode = lambda shape, dtype: (2178 * shape[2] * shape[3] * 64) * model_management.dtype_size(dtype) * VAE_KL_MEM_RATIO
         self.downscale_ratio = 8
         self.upscale_ratio = 8
         self.latent_channels = 4
@@ -291,6 +297,7 @@ class VAE:
         self.downscale_index_formula = None
         self.upscale_index_formula = None
         self.extra_1d_channel = None
+        self.crop_input = True
 
         if config is None:
             if "decoder.mid.block_1.mix_factor" in sd:
@@ -542,6 +549,25 @@ class VAE:
                 self.latent_channels = 3
                 self.latent_dim = 2
                 self.output_channels = 3
+            elif "vocoder.activation_post.downsample.lowpass.filter" in sd: #MMAudio VAE
+                sample_rate = 16000
+                if sample_rate == 16000:
+                    mode = '16k'
+                else:
+                    mode = '44k'
+
+                self.first_stage_model = comfy.ldm.mmaudio.vae.autoencoder.AudioAutoencoder(mode=mode)
+                self.memory_used_encode = lambda shape, dtype: (30 * shape[2]) * model_management.dtype_size(dtype)
+                self.memory_used_decode = lambda shape, dtype: (90 * shape[2] * 1411.2) * model_management.dtype_size(dtype)
+                self.latent_channels = 20
+                self.output_channels = 2
+                self.upscale_ratio = 512 * (44100 / sample_rate)
+                self.downscale_ratio = 512 * (44100 / sample_rate)
+                self.latent_dim = 1
+                self.process_output = lambda audio: audio
+                self.process_input = lambda audio: audio
+                self.working_dtypes = [torch.float32]
+                self.crop_input = False
             else:
                 logging.warning("WARNING: No VAE weights detected, VAE not initalized.")
                 self.first_stage_model = None
@@ -575,6 +601,9 @@ class VAE:
             raise RuntimeError("ERROR: VAE is invalid: None\n\nIf the VAE is from a checkpoint loader node your checkpoint does not contain a valid VAE.")
 
     def vae_encode_crop_pixels(self, pixels):
+        if not self.crop_input:
+            return pixels
+
         downscale_ratio = self.spacial_compression_encode()
 
         dims = pixels.shape[1:-1]
@@ -890,6 +919,7 @@ class TEModel(Enum):
     QWEN25_3B = 10
     QWEN25_7B = 11
     BYT5_SMALL_GLYPH = 12
+    GEMMA_3_4B = 13
 
 def detect_te_model(sd):
     if "text_model.encoder.layers.30.mlp.fc1.weight" in sd:
@@ -912,6 +942,8 @@ def detect_te_model(sd):
             return TEModel.BYT5_SMALL_GLYPH
         return TEModel.T5_BASE
     if 'model.layers.0.post_feedforward_layernorm.weight' in sd:
+        if 'model.layers.0.self_attn.q_norm.weight' in sd:
+            return TEModel.GEMMA_3_4B
         return TEModel.GEMMA_2_2B
     if 'model.layers.0.self_attn.k_proj.bias' in sd:
         weight = sd['model.layers.0.self_attn.k_proj.bias']
@@ -1016,6 +1048,10 @@ def load_text_encoder_state_dicts(state_dicts=[], embedding_directory=None, clip
             clip_target.clip = comfy.text_encoders.lumina2.te(**llama_detect(clip_data))
             clip_target.tokenizer = comfy.text_encoders.lumina2.LuminaTokenizer
             tokenizer_data["spiece_model"] = clip_data[0].get("spiece_model", None)
+        elif te_model == TEModel.GEMMA_3_4B:
+            clip_target.clip = comfy.text_encoders.lumina2.te(**llama_detect(clip_data), model_type="gemma3_4b")
+            clip_target.tokenizer = comfy.text_encoders.lumina2.NTokenizer
+            tokenizer_data["spiece_model"] = clip_data[0].get("spiece_model", None)
         elif te_model == TEModel.LLAMA3_8:
             clip_target.clip = comfy.text_encoders.hidream.hidream_clip(**llama_detect(clip_data),
                                                                         clip_l=False, clip_g=False, t5=False, llama=True, dtype_t5=None, t5xxl_scaled_fp8=None)

comfy/text_encoders/llama.py

@@ -3,6 +3,7 @@ import torch.nn as nn
 from dataclasses import dataclass
 from typing import Optional, Any
 import math
+import logging
 
 from comfy.ldm.modules.attention import optimized_attention_for_device
 import comfy.model_management
@@ -28,6 +29,9 @@ class Llama2Config:
     mlp_activation = "silu"
     qkv_bias = False
     rope_dims = None
+    q_norm = None
+    k_norm = None
+    rope_scale = None
 
 @dataclass
 class Qwen25_3BConfig:
@@ -46,6 +50,9 @@ class Qwen25_3BConfig:
     mlp_activation = "silu"
     qkv_bias = True
     rope_dims = None
+    q_norm = None
+    k_norm = None
+    rope_scale = None
 
 @dataclass
 class Qwen25_7BVLI_Config:
@@ -64,6 +71,9 @@ class Qwen25_7BVLI_Config:
     mlp_activation = "silu"
     qkv_bias = True
     rope_dims = [16, 24, 24]
+    q_norm = None
+    k_norm = None
+    rope_scale = None
 
 @dataclass
 class Gemma2_2B_Config:
@@ -82,6 +92,32 @@ class Gemma2_2B_Config:
     mlp_activation = "gelu_pytorch_tanh"
     qkv_bias = False
     rope_dims = None
+    q_norm = None
+    k_norm = None
+    sliding_attention = None
+    rope_scale = None
+
+@dataclass
+class Gemma3_4B_Config:
+    vocab_size: int = 262208
+    hidden_size: int = 2560
+    intermediate_size: int = 10240
+    num_hidden_layers: int = 34
+    num_attention_heads: int = 8
+    num_key_value_heads: int = 4
+    max_position_embeddings: int = 131072
+    rms_norm_eps: float = 1e-6
+    rope_theta = [10000.0, 1000000.0]
+    transformer_type: str = "gemma3"
+    head_dim = 256
+    rms_norm_add = True
+    mlp_activation = "gelu_pytorch_tanh"
+    qkv_bias = False
+    rope_dims = None
+    q_norm = "gemma3"
+    k_norm = "gemma3"
+    sliding_attention = [False, False, False, False, False, 1024]
+    rope_scale = [1.0, 8.0]
 
 class RMSNorm(nn.Module):
     def __init__(self, dim: int, eps: float = 1e-5, add=False, device=None, dtype=None):
@@ -106,25 +142,40 @@ def rotate_half(x):
     return torch.cat((-x2, x1), dim=-1)
 
 
-def precompute_freqs_cis(head_dim, position_ids, theta, rope_dims=None, device=None):
-    theta_numerator = torch.arange(0, head_dim, 2, device=device).float()
-    inv_freq = 1.0 / (theta ** (theta_numerator / head_dim))
+def precompute_freqs_cis(head_dim, position_ids, theta, rope_scale=None, rope_dims=None, device=None):
+    if not isinstance(theta, list):
+        theta = [theta]
 
-    inv_freq_expanded = inv_freq[None, :, None].float().expand(position_ids.shape[0], -1, 1)
-    position_ids_expanded = position_ids[:, None, :].float()
-    freqs = (inv_freq_expanded.float() @ position_ids_expanded.float()).transpose(1, 2)
-    emb = torch.cat((freqs, freqs), dim=-1)
-    cos = emb.cos()
-    sin = emb.sin()
-    if rope_dims is not None and position_ids.shape[0] > 1:
-        mrope_section = rope_dims * 2
-        cos = torch.cat([m[i % 3] for i, m in enumerate(cos.split(mrope_section, dim=-1))], dim=-1).unsqueeze(0)
-        sin = torch.cat([m[i % 3] for i, m in enumerate(sin.split(mrope_section, dim=-1))], dim=-1).unsqueeze(0)
-    else:
-        cos = cos.unsqueeze(1)
-        sin = sin.unsqueeze(1)
+    out = []
+    for index, t in enumerate(theta):
+        theta_numerator = torch.arange(0, head_dim, 2, device=device).float()
+        inv_freq = 1.0 / (t ** (theta_numerator / head_dim))
 
-    return (cos, sin)
+        if rope_scale is not None:
+            if isinstance(rope_scale, list):
+                inv_freq /= rope_scale[index]
+            else:
+                inv_freq /= rope_scale
+
+        inv_freq_expanded = inv_freq[None, :, None].float().expand(position_ids.shape[0], -1, 1)
+        position_ids_expanded = position_ids[:, None, :].float()
+        freqs = (inv_freq_expanded.float() @ position_ids_expanded.float()).transpose(1, 2)
+        emb = torch.cat((freqs, freqs), dim=-1)
+        cos = emb.cos()
+        sin = emb.sin()
+        if rope_dims is not None and position_ids.shape[0] > 1:
+            mrope_section = rope_dims * 2
+            cos = torch.cat([m[i % 3] for i, m in enumerate(cos.split(mrope_section, dim=-1))], dim=-1).unsqueeze(0)
+            sin = torch.cat([m[i % 3] for i, m in enumerate(sin.split(mrope_section, dim=-1))], dim=-1).unsqueeze(0)
+        else:
+            cos = cos.unsqueeze(1)
+            sin = sin.unsqueeze(1)
+        out.append((cos, sin))
+
+    if len(out) == 1:
+        return out[0]
+
+    return out
 
 
 def apply_rope(xq, xk, freqs_cis):
@@ -152,6 +203,14 @@ class Attention(nn.Module):
         self.v_proj = ops.Linear(config.hidden_size, self.num_kv_heads * self.head_dim, bias=config.qkv_bias, device=device, dtype=dtype)
         self.o_proj = ops.Linear(self.inner_size, config.hidden_size, bias=False, device=device, dtype=dtype)
 
+        self.q_norm = None
+        self.k_norm = None
+
+        if config.q_norm == "gemma3":
+            self.q_norm = RMSNorm(self.head_dim, eps=config.rms_norm_eps, add=config.rms_norm_add, device=device, dtype=dtype)
+        if config.k_norm == "gemma3":
+            self.k_norm = RMSNorm(self.head_dim, eps=config.rms_norm_eps, add=config.rms_norm_add, device=device, dtype=dtype)
+
     def forward(
         self,
         hidden_states: torch.Tensor,
@@ -168,6 +227,11 @@ class Attention(nn.Module):
         xk = xk.view(batch_size, seq_length, self.num_kv_heads, self.head_dim).transpose(1, 2)
         xv = xv.view(batch_size, seq_length, self.num_kv_heads, self.head_dim).transpose(1, 2)
 
+        if self.q_norm is not None:
+            xq = self.q_norm(xq)
+        if self.k_norm is not None:
+            xk = self.k_norm(xk)
+
         xq, xk = apply_rope(xq, xk, freqs_cis=freqs_cis)
 
         xk = xk.repeat_interleave(self.num_heads // self.num_kv_heads, dim=1)
@@ -192,7 +256,7 @@ class MLP(nn.Module):
         return self.down_proj(self.activation(self.gate_proj(x)) * self.up_proj(x))
 
 class TransformerBlock(nn.Module):
-    def __init__(self, config: Llama2Config, device=None, dtype=None, ops: Any = None):
+    def __init__(self, config: Llama2Config, index, device=None, dtype=None, ops: Any = None):
         super().__init__()
         self.self_attn = Attention(config, device=device, dtype=dtype, ops=ops)
         self.mlp = MLP(config, device=device, dtype=dtype, ops=ops)
@@ -226,7 +290,7 @@ class TransformerBlock(nn.Module):
         return x
 
 class TransformerBlockGemma2(nn.Module):
-    def __init__(self, config: Llama2Config, device=None, dtype=None, ops: Any = None):
+    def __init__(self, config: Llama2Config, index, device=None, dtype=None, ops: Any = None):
         super().__init__()
         self.self_attn = Attention(config, device=device, dtype=dtype, ops=ops)
         self.mlp = MLP(config, device=device, dtype=dtype, ops=ops)
@@ -235,6 +299,13 @@ class TransformerBlockGemma2(nn.Module):
         self.pre_feedforward_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps, add=config.rms_norm_add, device=device, dtype=dtype)
         self.post_feedforward_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps, add=config.rms_norm_add, device=device, dtype=dtype)
 
+        if config.sliding_attention is not None:  # TODO: implement. (Not that necessary since models are trained on less than 1024 tokens)
+            self.sliding_attention = config.sliding_attention[index % len(config.sliding_attention)]
+        else:
+            self.sliding_attention = False
+
+        self.transformer_type = config.transformer_type
+
     def forward(
         self,
         x: torch.Tensor,
@@ -242,6 +313,14 @@ class TransformerBlockGemma2(nn.Module):
         freqs_cis: Optional[torch.Tensor] = None,
         optimized_attention=None,
     ):
+        if self.transformer_type == 'gemma3':
+            if self.sliding_attention:
+                if x.shape[1] > self.sliding_attention:
+                    logging.warning("Warning: sliding attention not implemented, results may be incorrect")
+                freqs_cis = freqs_cis[1]
+            else:
+                freqs_cis = freqs_cis[0]
+
         # Self Attention
         residual = x
         x = self.input_layernorm(x)
@@ -276,7 +355,7 @@ class Llama2_(nn.Module):
             device=device,
             dtype=dtype
         )
-        if self.config.transformer_type == "gemma2":
+        if self.config.transformer_type == "gemma2" or self.config.transformer_type == "gemma3":
             transformer = TransformerBlockGemma2
             self.normalize_in = True
         else:
@@ -284,8 +363,8 @@ class Llama2_(nn.Module):
             self.normalize_in = False
 
         self.layers = nn.ModuleList([
-            transformer(config, device=device, dtype=dtype, ops=ops)
-            for _ in range(config.num_hidden_layers)
+            transformer(config, index=i, device=device, dtype=dtype, ops=ops)
+            for i in range(config.num_hidden_layers)
         ])
         self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps, add=config.rms_norm_add, device=device, dtype=dtype)
         # self.lm_head = ops.Linear(config.hidden_size, config.vocab_size, bias=False, device=device, dtype=dtype)
@@ -305,6 +384,7 @@ class Llama2_(nn.Module):
         freqs_cis = precompute_freqs_cis(self.config.head_dim,
                                          position_ids,
                                          self.config.rope_theta,
+                                         self.config.rope_scale,
                                          self.config.rope_dims,
                                          device=x.device)
 
@@ -433,3 +513,12 @@ class Gemma2_2B(BaseLlama, torch.nn.Module):
 
         self.model = Llama2_(config, device=device, dtype=dtype, ops=operations)
         self.dtype = dtype
+
+class Gemma3_4B(BaseLlama, torch.nn.Module):
+    def __init__(self, config_dict, dtype, device, operations):
+        super().__init__()
+        config = Gemma3_4B_Config(**config_dict)
+        self.num_layers = config.num_hidden_layers
+
+        self.model = Llama2_(config, device=device, dtype=dtype, ops=operations)
+        self.dtype = dtype

comfy/text_encoders/lumina2.py

@@ -11,23 +11,41 @@ class Gemma2BTokenizer(sd1_clip.SDTokenizer):
     def state_dict(self):
         return {"spiece_model": self.tokenizer.serialize_model()}
 
+class Gemma3_4BTokenizer(sd1_clip.SDTokenizer):
+    def __init__(self, embedding_directory=None, tokenizer_data={}):
+        tokenizer = tokenizer_data.get("spiece_model", None)
+        super().__init__(tokenizer, pad_with_end=False, embedding_size=2560, embedding_key='gemma3_4b', tokenizer_class=SPieceTokenizer, has_end_token=False, pad_to_max_length=False, max_length=99999999, min_length=1, tokenizer_args={"add_bos": True, "add_eos": False}, tokenizer_data=tokenizer_data)
+
+    def state_dict(self):
+        return {"spiece_model": self.tokenizer.serialize_model()}
 
 class LuminaTokenizer(sd1_clip.SD1Tokenizer):
     def __init__(self, embedding_directory=None, tokenizer_data={}):
         super().__init__(embedding_directory=embedding_directory, tokenizer_data=tokenizer_data, name="gemma2_2b", tokenizer=Gemma2BTokenizer)
 
+class NTokenizer(sd1_clip.SD1Tokenizer):
+    def __init__(self, embedding_directory=None, tokenizer_data={}):
+        super().__init__(embedding_directory=embedding_directory, tokenizer_data=tokenizer_data, name="gemma3_4b", tokenizer=Gemma3_4BTokenizer)
 
 class Gemma2_2BModel(sd1_clip.SDClipModel):
     def __init__(self, device="cpu", layer="hidden", layer_idx=-2, dtype=None, attention_mask=True, model_options={}):
         super().__init__(device=device, layer=layer, layer_idx=layer_idx, textmodel_json_config={}, dtype=dtype, special_tokens={"start": 2, "pad": 0}, layer_norm_hidden_state=False, model_class=comfy.text_encoders.llama.Gemma2_2B, enable_attention_masks=attention_mask, return_attention_masks=attention_mask, model_options=model_options)
 
+class Gemma3_4BModel(sd1_clip.SDClipModel):
+    def __init__(self, device="cpu", layer="hidden", layer_idx=-2, dtype=None, attention_mask=True, model_options={}):
+        super().__init__(device=device, layer=layer, layer_idx=layer_idx, textmodel_json_config={}, dtype=dtype, special_tokens={"start": 2, "pad": 0}, layer_norm_hidden_state=False, model_class=comfy.text_encoders.llama.Gemma3_4B, enable_attention_masks=attention_mask, return_attention_masks=attention_mask, model_options=model_options)
 
 class LuminaModel(sd1_clip.SD1ClipModel):
-    def __init__(self, device="cpu", dtype=None, model_options={}):
-        super().__init__(device=device, dtype=dtype, name="gemma2_2b", clip_model=Gemma2_2BModel, model_options=model_options)
+    def __init__(self, device="cpu", dtype=None, model_options={}, name="gemma2_2b", clip_model=Gemma2_2BModel):
+        super().__init__(device=device, dtype=dtype, name=name, clip_model=clip_model, model_options=model_options)
 
 
-def te(dtype_llama=None, llama_scaled_fp8=None):
+def te(dtype_llama=None, llama_scaled_fp8=None, model_type="gemma2_2b"):
+    if model_type == "gemma2_2b":
+        model = Gemma2_2BModel
+    elif model_type == "gemma3_4b":
+        model = Gemma3_4BModel
+
     class LuminaTEModel_(LuminaModel):
         def __init__(self, device="cpu", dtype=None, model_options={}):
             if llama_scaled_fp8 is not None and "scaled_fp8" not in model_options:
@@ -35,5 +53,5 @@ def te(dtype_llama=None, llama_scaled_fp8=None):
                 model_options["scaled_fp8"] = llama_scaled_fp8
             if dtype_llama is not None:
                 dtype = dtype_llama
-            super().__init__(device=device, dtype=dtype, model_options=model_options)
+            super().__init__(device=device, dtype=dtype, name=model_type, model_options=model_options, clip_model=model)
     return LuminaTEModel_

comfy/utils.py

@@ -39,7 +39,11 @@ if hasattr(torch.serialization, "add_safe_globals"):  # TODO: this was added in
         pass
     ModelCheckpoint.__module__ = "pytorch_lightning.callbacks.model_checkpoint"
 
-    from numpy.core.multiarray import scalar
+    def scalar(*args, **kwargs):
+        from numpy.core.multiarray import scalar as sc
+        return sc(*args, **kwargs)
+    scalar.__module__ = "numpy.core.multiarray"
+
     from numpy import dtype
     from numpy.dtypes import Float64DType
     from _codecs import encode

comfy_api/latest/__init__.py

@@ -8,8 +8,8 @@ from comfy_api.internal.async_to_sync import create_sync_class
 from comfy_api.latest._input import ImageInput, AudioInput, MaskInput, LatentInput, VideoInput
 from comfy_api.latest._input_impl import VideoFromFile, VideoFromComponents
 from comfy_api.latest._util import VideoCodec, VideoContainer, VideoComponents
-from comfy_api.latest._io import _IO as io  #noqa: F401
-from comfy_api.latest._ui import _UI as ui  #noqa: F401
+from . import _io as io
+from . import _ui as ui
 # from comfy_api.latest._resources import _RESOURCES as resources  #noqa: F401
 from comfy_execution.utils import get_executing_context
 from comfy_execution.progress import get_progress_state, PreviewImageTuple
@@ -114,6 +114,10 @@ if TYPE_CHECKING:
     ComfyAPISync: Type[comfy_api.latest.generated.ComfyAPISyncStub.ComfyAPISyncStub]
 ComfyAPISync = create_sync_class(ComfyAPI_latest)
 
+# create new aliases for io and ui
+IO = io
+UI = ui
+
 __all__ = [
     "ComfyAPI",
     "ComfyAPISync",
@@ -121,4 +125,8 @@ __all__ = [
     "InputImpl",
     "Types",
     "ComfyExtension",
+    "io",
+    "IO",
+    "ui",
+    "UI",
 ]

comfy_api/latest/_input/video_types.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 from abc import ABC, abstractmethod
-from typing import Optional, Union
+from typing import Optional, Union, IO
 import io
 import av
 from comfy_api.util import VideoContainer, VideoCodec, VideoComponents
@@ -23,7 +23,7 @@ class VideoInput(ABC):
     @abstractmethod
     def save_to(
         self,
-        path: str,
+        path: Union[str, IO[bytes]],
         format: VideoContainer = VideoContainer.AUTO,
         codec: VideoCodec = VideoCodec.AUTO,
         metadata: Optional[dict] = None

comfy_api/latest/_io.py

@@ -336,11 +336,25 @@ class Combo(ComfyTypeIO):
     class Input(WidgetInput):
         """Combo input (dropdown)."""
         Type = str
-        def __init__(self, id: str, options: list[str]=None, display_name: str=None, optional=False, tooltip: str=None, lazy: bool=None,
-                    default: str=None, control_after_generate: bool=None,
-                    upload: UploadType=None, image_folder: FolderType=None,
-                    remote: RemoteOptions=None,
-                    socketless: bool=None):
+        def __init__(
+            self,
+            id: str,
+            options: list[str] | list[int] | type[Enum] = None,
+            display_name: str=None,
+            optional=False,
+            tooltip: str=None,
+            lazy: bool=None,
+            default: str | int | Enum = None,
+            control_after_generate: bool=None,
+            upload: UploadType=None,
+            image_folder: FolderType=None,
+            remote: RemoteOptions=None,
+            socketless: bool=None,
+        ):
+            if isinstance(options, type) and issubclass(options, Enum):
+                options = [v.value for v in options]
+            if isinstance(default, Enum):
+                default = default.value
             super().__init__(id, display_name, optional, tooltip, lazy, default, socketless)
             self.multiselect = False
             self.options = options
@@ -1568,78 +1582,78 @@ class _UIOutput(ABC):
         ...
 
 
-class _IO:
-    FolderType = FolderType
-    UploadType = UploadType
-    RemoteOptions = RemoteOptions
-    NumberDisplay = NumberDisplay
+__all__ = [
+    "FolderType",
+    "UploadType",
+    "RemoteOptions",
+    "NumberDisplay",
 
-    comfytype = staticmethod(comfytype)
-    Custom = staticmethod(Custom)
-    Input = Input
-    WidgetInput = WidgetInput
-    Output = Output
-    ComfyTypeI = ComfyTypeI
-    ComfyTypeIO = ComfyTypeIO
-    #---------------------------------
+    "comfytype",
+    "Custom",
+    "Input",
+    "WidgetInput",
+    "Output",
+    "ComfyTypeI",
+    "ComfyTypeIO",
     # Supported Types
-    Boolean = Boolean
-    Int = Int
-    Float = Float
-    String = String
-    Combo = Combo
-    MultiCombo = MultiCombo
-    Image = Image
-    WanCameraEmbedding = WanCameraEmbedding
-    Webcam = Webcam
-    Mask = Mask
-    Latent = Latent
-    Conditioning = Conditioning
-    Sampler = Sampler
-    Sigmas = Sigmas
-    Noise = Noise
-    Guider = Guider
-    Clip = Clip
-    ControlNet = ControlNet
-    Vae = Vae
-    Model = Model
-    ClipVision = ClipVision
-    ClipVisionOutput = ClipVisionOutput
-    AudioEncoder = AudioEncoder
-    AudioEncoderOutput = AudioEncoderOutput
-    StyleModel = StyleModel
-    Gligen = Gligen
-    UpscaleModel = UpscaleModel
-    Audio = Audio
-    Video = Video
-    SVG = SVG
-    LoraModel = LoraModel
-    LossMap = LossMap
-    Voxel = Voxel
-    Mesh = Mesh
-    Hooks = Hooks
-    HookKeyframes = HookKeyframes
-    TimestepsRange = TimestepsRange
-    LatentOperation = LatentOperation
-    FlowControl = FlowControl
-    Accumulation = Accumulation
-    Load3DCamera = Load3DCamera
-    Load3D = Load3D
-    Load3DAnimation = Load3DAnimation
-    Photomaker = Photomaker
-    Point = Point
-    FaceAnalysis = FaceAnalysis
-    BBOX = BBOX
-    SEGS = SEGS
-    AnyType = AnyType
-    MultiType = MultiType
-    #---------------------------------
-    HiddenHolder = HiddenHolder
-    Hidden = Hidden
-    NodeInfoV1 = NodeInfoV1
-    NodeInfoV3 = NodeInfoV3
-    Schema = Schema
-    ComfyNode = ComfyNode
-    NodeOutput = NodeOutput
-    add_to_dict_v1 = staticmethod(add_to_dict_v1)
-    add_to_dict_v3 = staticmethod(add_to_dict_v3)
+    "Boolean",
+    "Int",
+    "Float",
+    "String",
+    "Combo",
+    "MultiCombo",
+    "Image",
+    "WanCameraEmbedding",
+    "Webcam",
+    "Mask",
+    "Latent",
+    "Conditioning",
+    "Sampler",
+    "Sigmas",
+    "Noise",
+    "Guider",
+    "Clip",
+    "ControlNet",
+    "Vae",
+    "Model",
+    "ClipVision",
+    "ClipVisionOutput",
+    "AudioEncoder",
+    "AudioEncoderOutput",
+    "StyleModel",
+    "Gligen",
+    "UpscaleModel",
+    "Audio",
+    "Video",
+    "SVG",
+    "LoraModel",
+    "LossMap",
+    "Voxel",
+    "Mesh",
+    "Hooks",
+    "HookKeyframes",
+    "TimestepsRange",
+    "LatentOperation",
+    "FlowControl",
+    "Accumulation",
+    "Load3DCamera",
+    "Load3D",
+    "Load3DAnimation",
+    "Photomaker",
+    "Point",
+    "FaceAnalysis",
+    "BBOX",
+    "SEGS",
+    "AnyType",
+    "MultiType",
+    # Other classes
+    "HiddenHolder",
+    "Hidden",
+    "NodeInfoV1",
+    "NodeInfoV3",
+    "Schema",
+    "ComfyNode",
+    "NodeOutput",
+    "add_to_dict_v1",
+    "add_to_dict_v3",
+]

comfy_api/latest/_ui.py

@@ -449,15 +449,16 @@ class PreviewText(_UIOutput):
         return {"text": (self.value,)}
 
 
-class _UI:
-    SavedResult = SavedResult
-    SavedImages = SavedImages
-    SavedAudios = SavedAudios
-    ImageSaveHelper = ImageSaveHelper
-    AudioSaveHelper = AudioSaveHelper
-    PreviewImage = PreviewImage
-    PreviewMask = PreviewMask
-    PreviewAudio = PreviewAudio
-    PreviewVideo = PreviewVideo
-    PreviewUI3D = PreviewUI3D
-    PreviewText = PreviewText
+__all__ = [
+    "SavedResult",
+    "SavedImages",
+    "SavedAudios",
+    "ImageSaveHelper",
+    "AudioSaveHelper",
+    "PreviewImage",
+    "PreviewMask",
+    "PreviewAudio",
+    "PreviewVideo",
+    "PreviewUI3D",
+    "PreviewText",
+]

comfy_api_nodes/apinode_utils.py

@@ -3,6 +3,7 @@ import aiohttp
 import io
 import logging
 import mimetypes
+import os
 from typing import Optional, Union
 from comfy.utils import common_upscale
 from comfy_api.input_impl import VideoFromFile
@@ -18,7 +19,7 @@ from comfy_api_nodes.apis.client import (
     UploadResponse,
 )
 from server import PromptServer
-
+from comfy.cli_args import args
 
 import numpy as np
 from PIL import Image
@@ -30,7 +31,9 @@ from io import BytesIO
 import av
 
 
-async def download_url_to_video_output(video_url: str, timeout: int = None) -> VideoFromFile:
+async def download_url_to_video_output(
+    video_url: str, timeout: int = None, auth_kwargs: Optional[dict[str, str]] = None
+) -> VideoFromFile:
     """Downloads a video from a URL and returns a `VIDEO` output.
 
     Args:
@@ -39,7 +42,7 @@ async def download_url_to_video_output(video_url: str, timeout: int = None) -> V
     Returns:
         A Comfy node `VIDEO` output.
     """
-    video_io = await download_url_to_bytesio(video_url, timeout)
+    video_io = await download_url_to_bytesio(video_url, timeout, auth_kwargs=auth_kwargs)
     if video_io is None:
         error_msg = f"Failed to download video from {video_url}"
         logging.error(error_msg)
@@ -152,7 +155,7 @@ def validate_aspect_ratio(
             raise TypeError(
                 f"Aspect ratio cannot reduce to any less than {minimum_ratio_str} ({minimum_ratio}), but was {aspect_ratio} ({calculated_ratio})."
             )
-        elif calculated_ratio > maximum_ratio:
+        if calculated_ratio > maximum_ratio:
             raise TypeError(
                 f"Aspect ratio cannot reduce to any greater than {maximum_ratio_str} ({maximum_ratio}), but was {aspect_ratio} ({calculated_ratio})."
             )
@@ -164,7 +167,9 @@ def mimetype_to_extension(mime_type: str) -> str:
     return mime_type.split("/")[-1].lower()
 
 
-async def download_url_to_bytesio(url: str, timeout: int = None) -> BytesIO:
+async def download_url_to_bytesio(
+    url: str, timeout: int = None, auth_kwargs: Optional[dict[str, str]] = None
+) -> BytesIO:
     """Downloads content from a URL using requests and returns it as BytesIO.
 
     Args:
@@ -174,9 +179,18 @@ async def download_url_to_bytesio(url: str, timeout: int = None) -> BytesIO:
     Returns:
         BytesIO object containing the downloaded content.
     """
+    headers = {}
+    if url.startswith("/proxy/"):
+        url = str(args.comfy_api_base).rstrip("/") + url
+        auth_token = auth_kwargs.get("auth_token")
+        comfy_api_key = auth_kwargs.get("comfy_api_key")
+        if auth_token:
+            headers["Authorization"] = f"Bearer {auth_token}"
+        elif comfy_api_key:
+            headers["X-API-KEY"] = comfy_api_key
     timeout_cfg = aiohttp.ClientTimeout(total=timeout) if timeout else None
     async with aiohttp.ClientSession(timeout=timeout_cfg) as session:
-        async with session.get(url) as resp:
+        async with session.get(url, headers=headers) as resp:
             resp.raise_for_status()  # Raises HTTPError for bad responses (4XX or 5XX)
             return BytesIO(await resp.read())
 
@@ -256,7 +270,7 @@ def tensor_to_bytesio(
         mime_type: Target image MIME type (e.g., 'image/png', 'image/jpeg', 'image/webp', 'video/mp4').
 
     Returns:
-        Named BytesIO object containing the image data.
+        Named BytesIO object containing the image data, with pointer set to the start of buffer.
     """
     if not mime_type:
         mime_type = "image/png"
@@ -418,7 +432,7 @@ async def upload_video_to_comfyapi(
                     f"Video duration ({actual_duration:.2f}s) exceeds the maximum allowed ({max_duration}s)."
                 )
         except Exception as e:
-            logging.error(f"Error getting video duration: {e}")
+            logging.error("Error getting video duration: %s", str(e))
             raise ValueError(f"Could not verify video duration from source: {e}") from e
 
     upload_mime_type = f"video/{container.value.lower()}"
@@ -689,3 +703,16 @@ def image_tensor_pair_to_batch(
             "center",
         ).movedim(1, -1)
     return torch.cat((image1, image2), dim=0)
+
+
+def get_size(path_or_object: Union[str, io.BytesIO]) -> int:
+    if isinstance(path_or_object, str):
+        return os.path.getsize(path_or_object)
+    return len(path_or_object.getvalue())
+
+
+def validate_container_format_is_mp4(video: VideoInput) -> None:
+    """Validates video container format is MP4."""
+    container_format = video.get_container_format()
+    if container_format not in ["mp4", "mov,mp4,m4a,3gp,3g2,mj2"]:
+        raise ValueError(f"Only MP4 container format supported. Got: {container_format}")

comfy_api_nodes/apis/client.py

@@ -98,7 +98,7 @@ import io
 import os
 import socket
 from aiohttp.client_exceptions import ClientError, ClientResponseError
-from typing import Dict, Type, Optional, Any, TypeVar, Generic, Callable, Tuple
+from typing import Type, Optional, Any, TypeVar, Generic, Callable
 from enum import Enum
 import json
 from urllib.parse import urljoin, urlparse
@@ -175,7 +175,7 @@ class ApiClient:
         max_retries: int = 3,
         retry_delay: float = 1.0,
         retry_backoff_factor: float = 2.0,
-        retry_status_codes: Optional[Tuple[int, ...]] = None,
+        retry_status_codes: Optional[tuple[int, ...]] = None,
         session: Optional[aiohttp.ClientSession] = None,
     ):
         self.base_url = base_url
@@ -199,9 +199,9 @@ class ApiClient:
 
     @staticmethod
     def _create_json_payload_args(
-        data: Optional[Dict[str, Any]] = None,
-        headers: Optional[Dict[str, str]] = None,
-    ) -> Dict[str, Any]:
+        data: Optional[dict[str, Any]] = None,
+        headers: Optional[dict[str, str]] = None,
+    ) -> dict[str, Any]:
         return {
             "json": data,
             "headers": headers,
@@ -209,24 +209,27 @@ class ApiClient:
 
     def _create_form_data_args(
         self,
-        data: Dict[str, Any] | None,
-        files: Dict[str, Any] | None,
-        headers: Optional[Dict[str, str]] = None,
+        data: dict[str, Any] | None,
+        files: dict[str, Any] | None,
+        headers: Optional[dict[str, str]] = None,
         multipart_parser: Callable | None = None,
-    ) -> Dict[str, Any]:
+    ) -> dict[str, Any]:
         if headers and "Content-Type" in headers:
             del headers["Content-Type"]
 
         if multipart_parser and data:
             data = multipart_parser(data)
 
-        form = aiohttp.FormData(default_to_multipart=True)
-        if data:  # regular text fields
-            for k, v in data.items():
-                if v is None:
-                    continue  # aiohttp fails to serialize "None" values
-                # aiohttp expects strings or bytes; convert enums etc.
-                form.add_field(k, str(v) if not isinstance(v, (bytes, bytearray)) else v)
+        if isinstance(data, aiohttp.FormData):
+            form = data  # If the parser already returned a FormData, pass it through
+        else:
+            form = aiohttp.FormData(default_to_multipart=True)
+            if data:  # regular text fields
+                for k, v in data.items():
+                    if v is None:
+                        continue  # aiohttp fails to serialize "None" values
+                    # aiohttp expects strings or bytes; convert enums etc.
+                    form.add_field(k, str(v) if not isinstance(v, (bytes, bytearray)) else v)
 
         if files:
             file_iter = files if isinstance(files, list) else files.items()
@@ -251,9 +254,9 @@ class ApiClient:
 
     @staticmethod
     def _create_urlencoded_form_data_args(
-        data: Dict[str, Any],
-        headers: Optional[Dict[str, str]] = None,
-    ) -> Dict[str, Any]:
+        data: dict[str, Any],
+        headers: Optional[dict[str, str]] = None,
+    ) -> dict[str, Any]:
         headers = headers or {}
         headers["Content-Type"] = "application/x-www-form-urlencoded"
         return {
@@ -261,7 +264,7 @@ class ApiClient:
             "headers": headers,
         }
 
-    def get_headers(self) -> Dict[str, str]:
+    def get_headers(self) -> dict[str, str]:
         """Get headers for API requests, including authentication if available"""
         headers = {"Content-Type": "application/json", "Accept": "application/json"}
 
@@ -272,7 +275,7 @@ class ApiClient:
 
         return headers
 
-    async def _check_connectivity(self, target_url: str) -> Dict[str, bool]:
+    async def _check_connectivity(self, target_url: str) -> dict[str, bool]:
         """
         Check connectivity to determine if network issues are local or server-related.
 
@@ -313,14 +316,14 @@ class ApiClient:
         self,
         method: str,
         path: str,
-        params: Optional[Dict[str, Any]] = None,
-        data: Optional[Dict[str, Any]] = None,
-        files: Optional[Dict[str, Any] | list[tuple[str, Any]]] = None,
-        headers: Optional[Dict[str, str]] = None,
+        params: Optional[dict[str, Any]] = None,
+        data: Optional[dict[str, Any]] = None,
+        files: Optional[dict[str, Any] | list[tuple[str, Any]]] = None,
+        headers: Optional[dict[str, str]] = None,
         content_type: str = "application/json",
         multipart_parser: Callable | None = None,
         retry_count: int = 0,  # Used internally for tracking retries
-    ) -> Dict[str, Any]:
+    ) -> dict[str, Any]:
         """
         Make an HTTP request to the API with automatic retries for transient errors.
 
@@ -356,10 +359,10 @@ class ApiClient:
         if params:
             params = {k: v for k, v in params.items() if v is not None}  # aiohttp fails to serialize None values
 
-        logging.debug(f"[DEBUG] Request Headers: {request_headers}")
-        logging.debug(f"[DEBUG] Files: {files}")
-        logging.debug(f"[DEBUG] Params: {params}")
-        logging.debug(f"[DEBUG] Data: {data}")
+        logging.debug("[DEBUG] Request Headers: %s", request_headers)
+        logging.debug("[DEBUG] Files: %s", files)
+        logging.debug("[DEBUG] Params: %s", params)
+        logging.debug("[DEBUG] Data: %s", data)
 
         if content_type == "application/x-www-form-urlencoded":
             payload_args = self._create_urlencoded_form_data_args(data or {}, request_headers)
@@ -482,7 +485,7 @@ class ApiClient:
             retry_delay: Initial delay between retries in seconds
             retry_backoff_factor: Multiplier for the delay after each retry
         """
-        headers: Dict[str, str] = {}
+        headers: dict[str, str] = {}
         skip_auto_headers: set[str] = set()
         if content_type:
             headers["Content-Type"] = content_type
@@ -555,7 +558,7 @@ class ApiClient:
         *req_meta,
         retry_count: int,
         response_content: dict | str = "",
-    ) -> Dict[str, Any]:
+    ) -> dict[str, Any]:
         status_code = exc.status
         if status_code == 401:
             user_friendly = "Unauthorized: Please login first to use this node."
@@ -589,9 +592,9 @@ class ApiClient:
             error_message=f"HTTP Error {exc.status}",
         )
 
-        logging.debug(f"[DEBUG] API Error: {user_friendly} (Status: {status_code})")
+        logging.debug("[DEBUG] API Error: %s (Status: %s)", user_friendly, status_code)
         if response_content:
-            logging.debug(f"[DEBUG] Response content: {response_content}")
+            logging.debug("[DEBUG] Response content: %s", response_content)
 
         # Retry if eligible
         if status_code in self.retry_status_codes and retry_count < self.max_retries:
@@ -656,7 +659,7 @@ class ApiEndpoint(Generic[T, R]):
         method: HttpMethod,
         request_model: Type[T],
         response_model: Type[R],
-        query_params: Optional[Dict[str, Any]] = None,
+        query_params: Optional[dict[str, Any]] = None,
     ):
         """Initialize an API endpoint definition.
 
@@ -681,11 +684,11 @@ class SynchronousOperation(Generic[T, R]):
         self,
         endpoint: ApiEndpoint[T, R],
         request: T,
-        files: Optional[Dict[str, Any] | list[tuple[str, Any]]] = None,
+        files: Optional[dict[str, Any] | list[tuple[str, Any]]] = None,
         api_base: str | None = None,
         auth_token: Optional[str] = None,
         comfy_api_key: Optional[str] = None,
-        auth_kwargs: Optional[Dict[str, str]] = None,
+        auth_kwargs: Optional[dict[str, str]] = None,
         timeout: float = 7200.0,
         verify_ssl: bool = True,
         content_type: str = "application/json",
@@ -726,7 +729,7 @@ class SynchronousOperation(Generic[T, R]):
             )
 
         try:
-            request_dict: Optional[Dict[str, Any]]
+            request_dict: Optional[dict[str, Any]]
             if isinstance(self.request, EmptyRequest):
                 request_dict = None
             else:
@@ -735,11 +738,9 @@ class SynchronousOperation(Generic[T, R]):
                     if isinstance(v, Enum):
                         request_dict[k] = v.value
 
-            logging.debug(
-                f"[DEBUG] API Request: {self.endpoint.method.value} {self.endpoint.path}"
-            )
-            logging.debug(f"[DEBUG] Request Data: {json.dumps(request_dict, indent=2)}")
-            logging.debug(f"[DEBUG] Query Params: {self.endpoint.query_params}")
+            logging.debug("[DEBUG] API Request: %s %s", self.endpoint.method.value, self.endpoint.path)
+            logging.debug("[DEBUG] Request Data: %s", json.dumps(request_dict, indent=2))
+            logging.debug("[DEBUG] Query Params: %s", self.endpoint.query_params)
 
             response_json = await client.request(
                 self.endpoint.method.value,
@@ -754,11 +755,11 @@ class SynchronousOperation(Generic[T, R]):
             logging.debug("=" * 50)
             logging.debug("[DEBUG] RESPONSE DETAILS:")
             logging.debug("[DEBUG] Status Code: 200 (Success)")
-            logging.debug(f"[DEBUG] Response Body: {json.dumps(response_json, indent=2)}")
+            logging.debug("[DEBUG] Response Body: %s", json.dumps(response_json, indent=2))
             logging.debug("=" * 50)
 
             parsed_response = self.endpoint.response_model.model_validate(response_json)
-            logging.debug(f"[DEBUG] Parsed Response: {parsed_response}")
+            logging.debug("[DEBUG] Parsed Response: %s", parsed_response)
             return parsed_response
         finally:
             if owns_client:
@@ -781,14 +782,16 @@ class PollingOperation(Generic[T, R]):
         poll_endpoint: ApiEndpoint[EmptyRequest, R],
         completed_statuses: list[str],
         failed_statuses: list[str],
-        status_extractor: Callable[[R], str],
-        progress_extractor: Callable[[R], float] | None = None,
-        result_url_extractor: Callable[[R], str] | None = None,
+        *,
+        status_extractor: Callable[[R], Optional[str]],
+        progress_extractor: Callable[[R], Optional[float]] | None = None,
+        result_url_extractor: Callable[[R], Optional[str]] | None = None,
+        price_extractor: Callable[[R], Optional[float]] | None = None,
         request: Optional[T] = None,
         api_base: str | None = None,
         auth_token: Optional[str] = None,
         comfy_api_key: Optional[str] = None,
-        auth_kwargs: Optional[Dict[str, str]] = None,
+        auth_kwargs: Optional[dict[str, str]] = None,
         poll_interval: float = 5.0,
         max_poll_attempts: int = 120,  # Default max polling attempts (10 minutes with 5s interval)
         max_retries: int = 3,  # Max retries per individual API call
@@ -814,10 +817,12 @@ class PollingOperation(Generic[T, R]):
         self.status_extractor = status_extractor or (lambda x: getattr(x, "status", None))
         self.progress_extractor = progress_extractor
         self.result_url_extractor = result_url_extractor
+        self.price_extractor = price_extractor
         self.node_id = node_id
         self.completed_statuses = completed_statuses
         self.failed_statuses = failed_statuses
         self.final_response: Optional[R] = None
+        self.extracted_price: Optional[float] = None
 
     async def execute(self, client: Optional[ApiClient] = None) -> R:
         owns_client = client is None
@@ -839,6 +844,8 @@ class PollingOperation(Generic[T, R]):
     def _display_text_on_node(self, text: str):
         if not self.node_id:
             return
+        if self.extracted_price is not None:
+            text = f"Price: ${self.extracted_price}\n{text}"
         PromptServer.instance.send_progress_text(text, self.node_id)
 
     def _display_time_progress_on_node(self, time_completed: int | float):
@@ -874,18 +881,19 @@ class PollingOperation(Generic[T, R]):
         status = TaskStatus.PENDING
         for poll_count in range(1, self.max_poll_attempts + 1):
             try:
-                logging.debug(f"[DEBUG] Polling attempt #{poll_count}")
+                logging.debug("[DEBUG] Polling attempt #%s", poll_count)
 
-                request_dict = (
-                    None if self.request is None else self.request.model_dump(exclude_none=True)
-                )
+                request_dict = None if self.request is None else self.request.model_dump(exclude_none=True)
 
                 if poll_count == 1:
                     logging.debug(
-                        f"[DEBUG] Poll Request: {self.poll_endpoint.method.value} {self.poll_endpoint.path}"
+                        "[DEBUG] Poll Request: %s %s",
+                        self.poll_endpoint.method.value,
+                        self.poll_endpoint.path,
                     )
                     logging.debug(
-                        f"[DEBUG] Poll Request Data: {json.dumps(request_dict, indent=2) if request_dict else 'None'}"
+                        "[DEBUG] Poll Request Data: %s",
+                        json.dumps(request_dict, indent=2) if request_dict else "None",
                     )
 
                 # Query task status
@@ -900,7 +908,7 @@ class PollingOperation(Generic[T, R]):
 
                 # Check if task is complete
                 status = self._check_task_status(response_obj)
-                logging.debug(f"[DEBUG] Task Status: {status}")
+                logging.debug("[DEBUG] Task Status: %s", status)
 
                 # If progress extractor is provided, extract progress
                 if self.progress_extractor:
@@ -908,13 +916,18 @@ class PollingOperation(Generic[T, R]):
                     if new_progress is not None:
                         progress.update_absolute(new_progress, total=PROGRESS_BAR_MAX)
 
+                if self.price_extractor:
+                    price = self.price_extractor(response_obj)
+                    if price is not None:
+                        self.extracted_price = price
+
                 if status == TaskStatus.COMPLETED:
                     message = "Task completed successfully"
                     if self.result_url_extractor:
                         result_url = self.result_url_extractor(response_obj)
                         if result_url:
                             message = f"Result URL: {result_url}"
-                    logging.debug(f"[DEBUG] {message}")
+                    logging.debug("[DEBUG] %s", message)
                     self._display_text_on_node(message)
                     self.final_response = response_obj
                     if self.progress_extractor:
@@ -922,7 +935,7 @@ class PollingOperation(Generic[T, R]):
                     return self.final_response
                 if status == TaskStatus.FAILED:
                     message = f"Task failed: {json.dumps(resp)}"
-                    logging.error(f"[DEBUG] {message}")
+                    logging.error("[DEBUG] %s", message)
                     raise Exception(message)
                 logging.debug("[DEBUG] Task still pending, continuing to poll...")
                 # Task pending – wait
@@ -936,7 +949,12 @@ class PollingOperation(Generic[T, R]):
                     raise Exception(
                         f"Polling aborted after {consecutive_errors} network errors: {str(e)}"
                     ) from e
-                logging.warning("Network error (%s/%s): %s", consecutive_errors, max_consecutive_errors, str(e))
+                logging.warning(
+                    "Network error (%s/%s): %s",
+                    consecutive_errors,
+                    max_consecutive_errors,
+                    str(e),
+                )
                 await asyncio.sleep(self.poll_interval)
             except Exception as e:
                 # For other errors, increment count and potentially abort
@@ -946,10 +964,13 @@ class PollingOperation(Generic[T, R]):
                         f"Polling aborted after {consecutive_errors} consecutive errors: {str(e)}"
                     ) from e
 
-                logging.error(f"[DEBUG] Polling error: {str(e)}")
+                logging.error("[DEBUG] Polling error: %s", str(e))
                 logging.warning(
-                    f"Error during polling (attempt {poll_count}/{self.max_poll_attempts}): {str(e)}. "
-                    f"Will retry in {self.poll_interval} seconds."
+                    "Error during polling (attempt %s/%s): %s. Will retry in %s seconds.",
+                    poll_count,
+                    self.max_poll_attempts,
+                    str(e),
+                    self.poll_interval,
                 )
                 await asyncio.sleep(self.poll_interval)
 

comfy_api_nodes/apis/gemini_api.py

@@ -1,19 +1,22 @@
-from __future__ import annotations
-
-from typing import List, Optional
+from typing import Optional
 
 from comfy_api_nodes.apis import GeminiGenerationConfig, GeminiContent, GeminiSafetySetting, GeminiSystemInstructionContent, GeminiTool, GeminiVideoMetadata
 from pydantic import BaseModel
 
 
+class GeminiImageConfig(BaseModel):
+    aspectRatio: Optional[str] = None
+
+
 class GeminiImageGenerationConfig(GeminiGenerationConfig):
-    responseModalities: Optional[List[str]] = None
+    responseModalities: Optional[list[str]] = None
+    imageConfig: Optional[GeminiImageConfig] = None
 
 
 class GeminiImageGenerateContentRequest(BaseModel):
-    contents: List[GeminiContent]
+    contents: list[GeminiContent]
     generationConfig: Optional[GeminiImageGenerationConfig] = None
-    safetySettings: Optional[List[GeminiSafetySetting]] = None
+    safetySettings: Optional[list[GeminiSafetySetting]] = None
     systemInstruction: Optional[GeminiSystemInstructionContent] = None
-    tools: Optional[List[GeminiTool]] = None
+    tools: Optional[list[GeminiTool]] = None
     videoMetadata: Optional[GeminiVideoMetadata] = None

comfy_api_nodes/apis/pika_defs.py

@@ -0,0 +1,100 @@
+from typing import Optional
+from enum import Enum
+from pydantic import BaseModel, Field
+
+
+class Pikaffect(str, Enum):
+    Cake_ify = "Cake-ify"
+    Crumble = "Crumble"
+    Crush = "Crush"
+    Decapitate = "Decapitate"
+    Deflate = "Deflate"
+    Dissolve = "Dissolve"
+    Explode = "Explode"
+    Eye_pop = "Eye-pop"
+    Inflate = "Inflate"
+    Levitate = "Levitate"
+    Melt = "Melt"
+    Peel = "Peel"
+    Poke = "Poke"
+    Squish = "Squish"
+    Ta_da = "Ta-da"
+    Tear = "Tear"
+
+
+class PikaBodyGenerate22C2vGenerate22PikascenesPost(BaseModel):
+    aspectRatio: Optional[float] = Field(None, description='Aspect ratio (width / height)')
+    duration: Optional[int] = Field(5)
+    ingredientsMode: str = Field(...)
+    negativePrompt: Optional[str] = Field(None)
+    promptText: Optional[str] = Field(None)
+    resolution: Optional[str] = Field('1080p')
+    seed: Optional[int] = Field(None)
+
+
+class PikaGenerateResponse(BaseModel):
+    video_id: str = Field(...)
+
+
+class PikaBodyGenerate22I2vGenerate22I2vPost(BaseModel):
+    duration: Optional[int] = 5
+    negativePrompt: Optional[str] = Field(None)
+    promptText: Optional[str] = Field(None)
+    resolution: Optional[str] = '1080p'
+    seed: Optional[int] = Field(None)
+
+
+class PikaBodyGenerate22KeyframeGenerate22PikaframesPost(BaseModel):
+    duration: Optional[int] = Field(None, ge=5, le=10)
+    negativePrompt: Optional[str] = Field(None)
+    promptText: str = Field(...)
+    resolution: Optional[str] = '1080p'
+    seed: Optional[int] = Field(None)
+
+
+class PikaBodyGenerate22T2vGenerate22T2vPost(BaseModel):
+    aspectRatio: Optional[float] = Field(
+        1.7777777777777777,
+        description='Aspect ratio (width / height)',
+        ge=0.4,
+        le=2.5,
+    )
+    duration: Optional[int] = 5
+    negativePrompt: Optional[str] = Field(None)
+    promptText: str = Field(...)
+    resolution: Optional[str] = '1080p'
+    seed: Optional[int] = Field(None)
+
+
+class PikaBodyGeneratePikadditionsGeneratePikadditionsPost(BaseModel):
+    negativePrompt: Optional[str] = Field(None)
+    promptText: Optional[str] = Field(None)
+    seed: Optional[int] = Field(None)
+
+
+class PikaBodyGeneratePikaffectsGeneratePikaffectsPost(BaseModel):
+    negativePrompt: Optional[str] = Field(None)
+    pikaffect: Optional[str] = None
+    promptText: Optional[str] = Field(None)
+    seed: Optional[int] = Field(None)
+
+
+class PikaBodyGeneratePikaswapsGeneratePikaswapsPost(BaseModel):
+    negativePrompt: Optional[str] = Field(None)
+    promptText: Optional[str] = Field(None)
+    seed: Optional[int] = Field(None)
+    modifyRegionRoi: Optional[str] = Field(None)
+
+
+class PikaStatusEnum(str, Enum):
+    queued = "queued"
+    started = "started"
+    finished = "finished"
+    failed = "failed"
+
+
+class PikaVideoResponse(BaseModel):
+    id: str = Field(...)
+    progress: Optional[int] = Field(None)
+    status: PikaStatusEnum
+    url: Optional[str] = Field(None)

comfy_api_nodes/apis/request_logger.py

@@ -21,7 +21,7 @@ def get_log_directory():
     try:
         os.makedirs(log_dir, exist_ok=True)
     except Exception as e:
-        logger.error(f"Error creating API log directory {log_dir}: {e}")
+        logger.error("Error creating API log directory %s: %s", log_dir, str(e))
         # Fallback to base temp directory if sub-directory creation fails
         return base_temp_dir
     return log_dir
@@ -122,9 +122,9 @@ def log_request_response(
     try:
         with open(filepath, "w", encoding="utf-8") as f:
             f.write("\n".join(log_content))
-        logger.debug(f"API log saved to: {filepath}")
+        logger.debug("API log saved to: %s", filepath)
     except Exception as e:
-        logger.error(f"Error writing API log to {filepath}: {e}")
+        logger.error("Error writing API log to %s: %s", filepath, str(e))
 
 
 if __name__ == '__main__':

comfy_api_nodes/nodes_bfl.py

@@ -3,7 +3,7 @@ import io
 from inspect import cleandoc
 from typing import Union, Optional
 from typing_extensions import override
-from comfy_api.latest import ComfyExtension, io as comfy_io
+from comfy_api.latest import ComfyExtension, IO
 from comfy_api_nodes.apis.bfl_api import (
     BFLStatus,
     BFLFluxExpandImageRequest,
@@ -131,7 +131,7 @@ def convert_image_to_base64(image: torch.Tensor):
     return base64.b64encode(img_byte_arr.getvalue()).decode()
 
 
-class FluxProUltraImageNode(comfy_io.ComfyNode):
+class FluxProUltraImageNode(IO.ComfyNode):
     """
     Generates images using Flux Pro 1.1 Ultra via api based on prompt and resolution.
     """
@@ -142,25 +142,25 @@ class FluxProUltraImageNode(comfy_io.ComfyNode):
     MAXIMUM_RATIO_STR = "4:1"
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="FluxProUltraImageNode",
             display_name="Flux 1.1 [pro] Ultra Image",
             category="api node/image/BFL",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the image generation",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "prompt_upsampling",
                     default=False,
                     tooltip="Whether to perform upsampling on the prompt. If active, automatically modifies the prompt for more creative generation, but results are nondeterministic (same seed will not produce exactly the same result).",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -168,21 +168,21 @@ class FluxProUltraImageNode(comfy_io.ComfyNode):
                     control_after_generate=True,
                     tooltip="The random seed used for creating the noise.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "aspect_ratio",
                     default="16:9",
                     tooltip="Aspect ratio of image; must be between 1:4 and 4:1.",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "raw",
                     default=False,
                     tooltip="When True, generate less processed, more natural-looking images.",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image_prompt",
                     optional=True,
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "image_prompt_strength",
                     default=0.1,
                     min=0.0,
@@ -192,11 +192,11 @@ class FluxProUltraImageNode(comfy_io.ComfyNode):
                     optional=True,
                 ),
             ],
-            outputs=[comfy_io.Image.Output()],
+            outputs=[IO.Image.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -225,7 +225,7 @@ class FluxProUltraImageNode(comfy_io.ComfyNode):
         seed=0,
         image_prompt=None,
         image_prompt_strength=0.1,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         if image_prompt is None:
             validate_string(prompt, strip_whitespace=False)
         operation = SynchronousOperation(
@@ -262,10 +262,10 @@ class FluxProUltraImageNode(comfy_io.ComfyNode):
             },
         )
         output_image = await handle_bfl_synchronous_operation(operation, node_id=cls.hidden.unique_id)
-        return comfy_io.NodeOutput(output_image)
+        return IO.NodeOutput(output_image)
 
 
-class FluxKontextProImageNode(comfy_io.ComfyNode):
+class FluxKontextProImageNode(IO.ComfyNode):
     """
     Edits images using Flux.1 Kontext [pro] via api based on prompt and aspect ratio.
     """
@@ -276,25 +276,25 @@ class FluxKontextProImageNode(comfy_io.ComfyNode):
     MAXIMUM_RATIO_STR = "4:1"
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id=cls.NODE_ID,
             display_name=cls.DISPLAY_NAME,
             category="api node/image/BFL",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the image generation - specify what and how to edit.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "aspect_ratio",
                     default="16:9",
                     tooltip="Aspect ratio of image; must be between 1:4 and 4:1.",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "guidance",
                     default=3.0,
                     min=0.1,
@@ -302,14 +302,14 @@ class FluxKontextProImageNode(comfy_io.ComfyNode):
                     step=0.1,
                     tooltip="Guidance strength for the image generation process",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "steps",
                     default=50,
                     min=1,
                     max=150,
                     tooltip="Number of steps for the image generation process",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=1234,
                     min=0,
@@ -317,21 +317,21 @@ class FluxKontextProImageNode(comfy_io.ComfyNode):
                     control_after_generate=True,
                     tooltip="The random seed used for creating the noise.",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "prompt_upsampling",
                     default=False,
                     tooltip="Whether to perform upsampling on the prompt. If active, automatically modifies the prompt for more creative generation, but results are nondeterministic (same seed will not produce exactly the same result).",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "input_image",
                     optional=True,
                 ),
             ],
-            outputs=[comfy_io.Image.Output()],
+            outputs=[IO.Image.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -350,7 +350,7 @@ class FluxKontextProImageNode(comfy_io.ComfyNode):
         input_image: Optional[torch.Tensor]=None,
         seed=0,
         prompt_upsampling=False,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         aspect_ratio = validate_aspect_ratio(
             aspect_ratio,
             minimum_ratio=cls.MINIMUM_RATIO,
@@ -386,7 +386,7 @@ class FluxKontextProImageNode(comfy_io.ComfyNode):
             },
         )
         output_image = await handle_bfl_synchronous_operation(operation, node_id=cls.hidden.unique_id)
-        return comfy_io.NodeOutput(output_image)
+        return IO.NodeOutput(output_image)
 
 
 class FluxKontextMaxImageNode(FluxKontextProImageNode):
@@ -400,45 +400,45 @@ class FluxKontextMaxImageNode(FluxKontextProImageNode):
     DISPLAY_NAME = "Flux.1 Kontext [max] Image"
 
 
-class FluxProImageNode(comfy_io.ComfyNode):
+class FluxProImageNode(IO.ComfyNode):
     """
     Generates images synchronously based on prompt and resolution.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="FluxProImageNode",
             display_name="Flux 1.1 [pro] Image",
             category="api node/image/BFL",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the image generation",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "prompt_upsampling",
                     default=False,
                     tooltip="Whether to perform upsampling on the prompt. If active, automatically modifies the prompt for more creative generation, but results are nondeterministic (same seed will not produce exactly the same result).",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "width",
                     default=1024,
                     min=256,
                     max=1440,
                     step=32,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "height",
                     default=768,
                     min=256,
                     max=1440,
                     step=32,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -446,7 +446,7 @@ class FluxProImageNode(comfy_io.ComfyNode):
                     control_after_generate=True,
                     tooltip="The random seed used for creating the noise.",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image_prompt",
                     optional=True,
                 ),
@@ -461,11 +461,11 @@ class FluxProImageNode(comfy_io.ComfyNode):
                 #     },
                 # ),
             ],
-            outputs=[comfy_io.Image.Output()],
+            outputs=[IO.Image.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -480,7 +480,7 @@ class FluxProImageNode(comfy_io.ComfyNode):
         seed=0,
         image_prompt=None,
         # image_prompt_strength=0.1,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         image_prompt = (
                     image_prompt
                     if image_prompt is None
@@ -508,77 +508,77 @@ class FluxProImageNode(comfy_io.ComfyNode):
             },
         )
         output_image = await handle_bfl_synchronous_operation(operation, node_id=cls.hidden.unique_id)
-        return comfy_io.NodeOutput(output_image)
+        return IO.NodeOutput(output_image)
 
 
-class FluxProExpandNode(comfy_io.ComfyNode):
+class FluxProExpandNode(IO.ComfyNode):
     """
     Outpaints image based on prompt.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="FluxProExpandNode",
             display_name="Flux.1 Expand Image",
             category="api node/image/BFL",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("image"),
-                comfy_io.String.Input(
+                IO.Image.Input("image"),
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the image generation",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "prompt_upsampling",
                     default=False,
                     tooltip="Whether to perform upsampling on the prompt. If active, automatically modifies the prompt for more creative generation, but results are nondeterministic (same seed will not produce exactly the same result).",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "top",
                     default=0,
                     min=0,
                     max=2048,
                     tooltip="Number of pixels to expand at the top of the image",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "bottom",
                     default=0,
                     min=0,
                     max=2048,
                     tooltip="Number of pixels to expand at the bottom of the image",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "left",
                     default=0,
                     min=0,
                     max=2048,
                     tooltip="Number of pixels to expand at the left of the image",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "right",
                     default=0,
                     min=0,
                     max=2048,
                     tooltip="Number of pixels to expand at the right of the image",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "guidance",
                     default=60,
                     min=1.5,
                     max=100,
                     tooltip="Guidance strength for the image generation process",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "steps",
                     default=50,
                     min=15,
                     max=50,
                     tooltip="Number of steps for the image generation process",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -587,11 +587,11 @@ class FluxProExpandNode(comfy_io.ComfyNode):
                     tooltip="The random seed used for creating the noise.",
                 ),
             ],
-            outputs=[comfy_io.Image.Output()],
+            outputs=[IO.Image.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -609,7 +609,7 @@ class FluxProExpandNode(comfy_io.ComfyNode):
         steps: int,
         guidance: float,
         seed=0,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         image = convert_image_to_base64(image)
 
         operation = SynchronousOperation(
@@ -637,51 +637,51 @@ class FluxProExpandNode(comfy_io.ComfyNode):
             },
         )
         output_image = await handle_bfl_synchronous_operation(operation, node_id=cls.hidden.unique_id)
-        return comfy_io.NodeOutput(output_image)
+        return IO.NodeOutput(output_image)
 
 
 
-class FluxProFillNode(comfy_io.ComfyNode):
+class FluxProFillNode(IO.ComfyNode):
     """
     Inpaints image based on mask and prompt.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="FluxProFillNode",
             display_name="Flux.1 Fill Image",
             category="api node/image/BFL",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("image"),
-                comfy_io.Mask.Input("mask"),
-                comfy_io.String.Input(
+                IO.Image.Input("image"),
+                IO.Mask.Input("mask"),
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the image generation",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "prompt_upsampling",
                     default=False,
                     tooltip="Whether to perform upsampling on the prompt. If active, automatically modifies the prompt for more creative generation, but results are nondeterministic (same seed will not produce exactly the same result).",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "guidance",
                     default=60,
                     min=1.5,
                     max=100,
                     tooltip="Guidance strength for the image generation process",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "steps",
                     default=50,
                     min=15,
                     max=50,
                     tooltip="Number of steps for the image generation process",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -690,11 +690,11 @@ class FluxProFillNode(comfy_io.ComfyNode):
                     tooltip="The random seed used for creating the noise.",
                 ),
             ],
-            outputs=[comfy_io.Image.Output()],
+            outputs=[IO.Image.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -709,7 +709,7 @@ class FluxProFillNode(comfy_io.ComfyNode):
         steps: int,
         guidance: float,
         seed=0,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         # prepare mask
         mask = resize_mask_to_image(mask, image)
         mask = convert_image_to_base64(convert_mask_to_image(mask))
@@ -738,35 +738,35 @@ class FluxProFillNode(comfy_io.ComfyNode):
             },
         )
         output_image = await handle_bfl_synchronous_operation(operation, node_id=cls.hidden.unique_id)
-        return comfy_io.NodeOutput(output_image)
+        return IO.NodeOutput(output_image)
 
 
-class FluxProCannyNode(comfy_io.ComfyNode):
+class FluxProCannyNode(IO.ComfyNode):
     """
     Generate image using a control image (canny).
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="FluxProCannyNode",
             display_name="Flux.1 Canny Control Image",
             category="api node/image/BFL",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("control_image"),
-                comfy_io.String.Input(
+                IO.Image.Input("control_image"),
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the image generation",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "prompt_upsampling",
                     default=False,
                     tooltip="Whether to perform upsampling on the prompt. If active, automatically modifies the prompt for more creative generation, but results are nondeterministic (same seed will not produce exactly the same result).",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "canny_low_threshold",
                     default=0.1,
                     min=0.01,
@@ -774,7 +774,7 @@ class FluxProCannyNode(comfy_io.ComfyNode):
                     step=0.01,
                     tooltip="Low threshold for Canny edge detection; ignored if skip_processing is True",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "canny_high_threshold",
                     default=0.4,
                     min=0.01,
@@ -782,26 +782,26 @@ class FluxProCannyNode(comfy_io.ComfyNode):
                     step=0.01,
                     tooltip="High threshold for Canny edge detection; ignored if skip_processing is True",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "skip_preprocessing",
                     default=False,
                     tooltip="Whether to skip preprocessing; set to True if control_image already is canny-fied, False if it is a raw image.",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "guidance",
                     default=30,
                     min=1,
                     max=100,
                     tooltip="Guidance strength for the image generation process",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "steps",
                     default=50,
                     min=15,
                     max=50,
                     tooltip="Number of steps for the image generation process",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -810,11 +810,11 @@ class FluxProCannyNode(comfy_io.ComfyNode):
                     tooltip="The random seed used for creating the noise.",
                 ),
             ],
-            outputs=[comfy_io.Image.Output()],
+            outputs=[IO.Image.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -831,7 +831,7 @@ class FluxProCannyNode(comfy_io.ComfyNode):
         steps: int,
         guidance: float,
         seed=0,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         control_image = convert_image_to_base64(control_image[:, :, :, :3])
         preprocessed_image = None
 
@@ -872,54 +872,54 @@ class FluxProCannyNode(comfy_io.ComfyNode):
             },
         )
         output_image = await handle_bfl_synchronous_operation(operation, node_id=cls.hidden.unique_id)
-        return comfy_io.NodeOutput(output_image)
+        return IO.NodeOutput(output_image)
 
 
-class FluxProDepthNode(comfy_io.ComfyNode):
+class FluxProDepthNode(IO.ComfyNode):
     """
     Generate image using a control image (depth).
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="FluxProDepthNode",
             display_name="Flux.1 Depth Control Image",
             category="api node/image/BFL",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("control_image"),
-                comfy_io.String.Input(
+                IO.Image.Input("control_image"),
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the image generation",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "prompt_upsampling",
                     default=False,
                     tooltip="Whether to perform upsampling on the prompt. If active, automatically modifies the prompt for more creative generation, but results are nondeterministic (same seed will not produce exactly the same result).",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "skip_preprocessing",
                     default=False,
                     tooltip="Whether to skip preprocessing; set to True if control_image already is depth-ified, False if it is a raw image.",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "guidance",
                     default=15,
                     min=1,
                     max=100,
                     tooltip="Guidance strength for the image generation process",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "steps",
                     default=50,
                     min=15,
                     max=50,
                     tooltip="Number of steps for the image generation process",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -928,11 +928,11 @@ class FluxProDepthNode(comfy_io.ComfyNode):
                     tooltip="The random seed used for creating the noise.",
                 ),
             ],
-            outputs=[comfy_io.Image.Output()],
+            outputs=[IO.Image.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -947,7 +947,7 @@ class FluxProDepthNode(comfy_io.ComfyNode):
         steps: int,
         guidance: float,
         seed=0,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         control_image = convert_image_to_base64(control_image[:,:,:,:3])
         preprocessed_image = None
 
@@ -977,12 +977,12 @@ class FluxProDepthNode(comfy_io.ComfyNode):
             },
         )
         output_image = await handle_bfl_synchronous_operation(operation, node_id=cls.hidden.unique_id)
-        return comfy_io.NodeOutput(output_image)
+        return IO.NodeOutput(output_image)
 
 
 class BFLExtension(ComfyExtension):
     @override
-    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
         return [
             FluxProUltraImageNode,
             # FluxProImageNode,

comfy_api_nodes/nodes_bytedance.py

@@ -7,7 +7,7 @@ from typing_extensions import override
 import torch
 from pydantic import BaseModel, Field
 
-from comfy_api.latest import ComfyExtension, io as comfy_io
+from comfy_api.latest import ComfyExtension, IO
 from comfy_api_nodes.util.validation_utils import (
     validate_image_aspect_ratio_range,
     get_number_of_images,
@@ -237,33 +237,33 @@ async def poll_until_finished(
     ).execute()
 
 
-class ByteDanceImageNode(comfy_io.ComfyNode):
+class ByteDanceImageNode(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="ByteDanceImageNode",
             display_name="ByteDance Image",
             category="api node/image/ByteDance",
             description="Generate images using ByteDance models via api based on prompt",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
-                    options=[model.value for model in Text2ImageModelName],
-                    default=Text2ImageModelName.seedream_3.value,
+                    options=Text2ImageModelName,
+                    default=Text2ImageModelName.seedream_3,
                     tooltip="Model name",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     tooltip="The text prompt used to generate the image",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "size_preset",
                     options=[label for label, _, _ in RECOMMENDED_PRESETS],
                     tooltip="Pick a recommended size. Select Custom to use the width and height below",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "width",
                     default=1024,
                     min=512,
@@ -271,7 +271,7 @@ class ByteDanceImageNode(comfy_io.ComfyNode):
                     step=64,
                     tooltip="Custom width for image. Value is working only if `size_preset` is set to `Custom`",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "height",
                     default=1024,
                     min=512,
@@ -279,28 +279,28 @@ class ByteDanceImageNode(comfy_io.ComfyNode):
                     step=64,
                     tooltip="Custom height for image. Value is working only if `size_preset` is set to `Custom`",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed to use for generation",
                     optional=True,
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "guidance_scale",
                     default=2.5,
                     min=1.0,
                     max=10.0,
                     step=0.01,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Higher value makes the image follow the prompt more closely",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "watermark",
                     default=True,
                     tooltip="Whether to add an \"AI generated\" watermark to the image",
@@ -308,12 +308,12 @@ class ByteDanceImageNode(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -329,7 +329,7 @@ class ByteDanceImageNode(comfy_io.ComfyNode):
         seed: int,
         guidance_scale: float,
         watermark: bool,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=True, min_length=1)
         w = h = None
         for label, tw, th in RECOMMENDED_PRESETS:
@@ -367,57 +367,57 @@ class ByteDanceImageNode(comfy_io.ComfyNode):
             request=payload,
             auth_kwargs=auth_kwargs,
         ).execute()
-        return comfy_io.NodeOutput(await download_url_to_image_tensor(get_image_url_from_response(response)))
+        return IO.NodeOutput(await download_url_to_image_tensor(get_image_url_from_response(response)))
 
 
-class ByteDanceImageEditNode(comfy_io.ComfyNode):
+class ByteDanceImageEditNode(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="ByteDanceImageEditNode",
             display_name="ByteDance Image Edit",
             category="api node/image/ByteDance",
             description="Edit images using ByteDance models via api based on prompt",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
-                    options=[model.value for model in Image2ImageModelName],
-                    default=Image2ImageModelName.seededit_3.value,
+                    options=Image2ImageModelName,
+                    default=Image2ImageModelName.seededit_3,
                     tooltip="Model name",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                     tooltip="The base image to edit",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Instruction to edit image",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed to use for generation",
                     optional=True,
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "guidance_scale",
                     default=5.5,
                     min=1.0,
                     max=10.0,
                     step=0.01,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Higher value makes the image follow the prompt more closely",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "watermark",
                     default=True,
                     tooltip="Whether to add an \"AI generated\" watermark to the image",
@@ -425,12 +425,12 @@ class ByteDanceImageEditNode(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -444,7 +444,7 @@ class ByteDanceImageEditNode(comfy_io.ComfyNode):
         seed: int,
         guidance_scale: float,
         watermark: bool,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=True, min_length=1)
         if get_number_of_images(image) != 1:
             raise ValueError("Exactly one input image is required.")
@@ -477,42 +477,42 @@ class ByteDanceImageEditNode(comfy_io.ComfyNode):
             request=payload,
             auth_kwargs=auth_kwargs,
         ).execute()
-        return comfy_io.NodeOutput(await download_url_to_image_tensor(get_image_url_from_response(response)))
+        return IO.NodeOutput(await download_url_to_image_tensor(get_image_url_from_response(response)))
 
 
-class ByteDanceSeedreamNode(comfy_io.ComfyNode):
+class ByteDanceSeedreamNode(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="ByteDanceSeedreamNode",
             display_name="ByteDance Seedream 4",
             category="api node/image/ByteDance",
             description="Unified text-to-image generation and precise single-sentence editing at up to 4K resolution.",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=["seedream-4-0-250828"],
                     tooltip="Model name",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Text prompt for creating or editing an image.",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                     tooltip="Input image(s) for image-to-image generation. "
                             "List of 1-10 images for single or multi-reference generation.",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "size_preset",
                     options=[label for label, _, _ in RECOMMENDED_PRESETS_SEEDREAM_4],
                     tooltip="Pick a recommended size. Select Custom to use the width and height below.",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "width",
                     default=2048,
                     min=1024,
@@ -521,7 +521,7 @@ class ByteDanceSeedreamNode(comfy_io.ComfyNode):
                     tooltip="Custom width for image. Value is working only if `size_preset` is set to `Custom`",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "height",
                     default=2048,
                     min=1024,
@@ -530,7 +530,7 @@ class ByteDanceSeedreamNode(comfy_io.ComfyNode):
                     tooltip="Custom height for image. Value is working only if `size_preset` is set to `Custom`",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "sequential_image_generation",
                     options=["disabled", "auto"],
                     tooltip="Group image generation mode. "
@@ -539,35 +539,35 @@ class ByteDanceSeedreamNode(comfy_io.ComfyNode):
                             "(e.g., story scenes, character variations).",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "max_images",
                     default=1,
                     min=1,
                     max=15,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Maximum number of images to generate when sequential_image_generation='auto'. "
                             "Total images (input + generated) cannot exceed 15.",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed to use for generation.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "watermark",
                     default=True,
                     tooltip="Whether to add an \"AI generated\" watermark to the image.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "fail_on_partial",
                     default=True,
                     tooltip="If enabled, abort execution if any requested images are missing or return an error.",
@@ -575,12 +575,12 @@ class ByteDanceSeedreamNode(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -599,7 +599,7 @@ class ByteDanceSeedreamNode(comfy_io.ComfyNode):
         seed: int = 0,
         watermark: bool = True,
         fail_on_partial: bool = True,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=True, min_length=1)
         w = h = None
         for label, tw, th in RECOMMENDED_PRESETS_SEEDREAM_4:
@@ -657,72 +657,72 @@ class ByteDanceSeedreamNode(comfy_io.ComfyNode):
         ).execute()
 
         if len(response.data) == 1:
-            return comfy_io.NodeOutput(await download_url_to_image_tensor(get_image_url_from_response(response)))
+            return IO.NodeOutput(await download_url_to_image_tensor(get_image_url_from_response(response)))
         urls = [str(d["url"]) for d in response.data if isinstance(d, dict) and "url" in d]
         if fail_on_partial and len(urls) < len(response.data):
             raise RuntimeError(f"Only {len(urls)} of {len(response.data)} images were generated before error.")
-        return comfy_io.NodeOutput(torch.cat([await download_url_to_image_tensor(i) for i in urls]))
+        return IO.NodeOutput(torch.cat([await download_url_to_image_tensor(i) for i in urls]))
 
 
-class ByteDanceTextToVideoNode(comfy_io.ComfyNode):
+class ByteDanceTextToVideoNode(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="ByteDanceTextToVideoNode",
             display_name="ByteDance Text to Video",
             category="api node/video/ByteDance",
             description="Generate video using ByteDance models via api based on prompt",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
-                    options=[model.value for model in Text2VideoModelName],
-                    default=Text2VideoModelName.seedance_1_pro.value,
+                    options=Text2VideoModelName,
+                    default=Text2VideoModelName.seedance_1_pro,
                     tooltip="Model name",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     tooltip="The text prompt used to generate the video.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
                     options=["480p", "720p", "1080p"],
                     tooltip="The resolution of the output video.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
                     options=["16:9", "4:3", "1:1", "3:4", "9:16", "21:9"],
                     tooltip="The aspect ratio of the output video.",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "duration",
                     default=5,
                     min=3,
                     max=12,
                     step=1,
                     tooltip="The duration of the output video in seconds.",
-                    display_mode=comfy_io.NumberDisplay.slider,
+                    display_mode=IO.NumberDisplay.slider,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed to use for generation.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "camera_fixed",
                     default=False,
                     tooltip="Specifies whether to fix the camera. The platform appends an instruction "
                             "to fix the camera to your prompt, but does not guarantee the actual effect.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "watermark",
                     default=True,
                     tooltip="Whether to add an \"AI generated\" watermark to the video.",
@@ -730,12 +730,12 @@ class ByteDanceTextToVideoNode(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -751,7 +751,7 @@ class ByteDanceTextToVideoNode(comfy_io.ComfyNode):
         seed: int,
         camera_fixed: bool,
         watermark: bool,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=True, min_length=1)
         raise_if_text_params(prompt, ["resolution", "ratio", "duration", "seed", "camerafixed", "watermark"])
 
@@ -781,69 +781,69 @@ class ByteDanceTextToVideoNode(comfy_io.ComfyNode):
         )
 
 
-class ByteDanceImageToVideoNode(comfy_io.ComfyNode):
+class ByteDanceImageToVideoNode(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="ByteDanceImageToVideoNode",
             display_name="ByteDance Image to Video",
             category="api node/video/ByteDance",
             description="Generate video using ByteDance models via api based on image and prompt",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
-                    options=[model.value for model in Image2VideoModelName],
-                    default=Image2VideoModelName.seedance_1_pro.value,
+                    options=Image2VideoModelName,
+                    default=Image2VideoModelName.seedance_1_pro,
                     tooltip="Model name",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     tooltip="The text prompt used to generate the video.",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                     tooltip="First frame to be used for the video.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
                     options=["480p", "720p", "1080p"],
                     tooltip="The resolution of the output video.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
                     options=["adaptive", "16:9", "4:3", "1:1", "3:4", "9:16", "21:9"],
                     tooltip="The aspect ratio of the output video.",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "duration",
                     default=5,
                     min=3,
                     max=12,
                     step=1,
                     tooltip="The duration of the output video in seconds.",
-                    display_mode=comfy_io.NumberDisplay.slider,
+                    display_mode=IO.NumberDisplay.slider,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed to use for generation.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "camera_fixed",
                     default=False,
                     tooltip="Specifies whether to fix the camera. The platform appends an instruction "
                             "to fix the camera to your prompt, but does not guarantee the actual effect.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "watermark",
                     default=True,
                     tooltip="Whether to add an \"AI generated\" watermark to the video.",
@@ -851,12 +851,12 @@ class ByteDanceImageToVideoNode(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -873,7 +873,7 @@ class ByteDanceImageToVideoNode(comfy_io.ComfyNode):
         seed: int,
         camera_fixed: bool,
         watermark: bool,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=True, min_length=1)
         raise_if_text_params(prompt, ["resolution", "ratio", "duration", "seed", "camerafixed", "watermark"])
         validate_image_dimensions(image, min_width=300, min_height=300, max_width=6000, max_height=6000)
@@ -908,73 +908,73 @@ class ByteDanceImageToVideoNode(comfy_io.ComfyNode):
         )
 
 
-class ByteDanceFirstLastFrameNode(comfy_io.ComfyNode):
+class ByteDanceFirstLastFrameNode(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="ByteDanceFirstLastFrameNode",
             display_name="ByteDance First-Last-Frame to Video",
             category="api node/video/ByteDance",
             description="Generate video using prompt and first and last frames.",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=[model.value for model in Image2VideoModelName],
                     default=Image2VideoModelName.seedance_1_lite.value,
                     tooltip="Model name",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     tooltip="The text prompt used to generate the video.",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "first_frame",
                     tooltip="First frame to be used for the video.",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "last_frame",
                     tooltip="Last frame to be used for the video.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
                     options=["480p", "720p", "1080p"],
                     tooltip="The resolution of the output video.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
                     options=["adaptive", "16:9", "4:3", "1:1", "3:4", "9:16", "21:9"],
                     tooltip="The aspect ratio of the output video.",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "duration",
                     default=5,
                     min=3,
                     max=12,
                     step=1,
                     tooltip="The duration of the output video in seconds.",
-                    display_mode=comfy_io.NumberDisplay.slider,
+                    display_mode=IO.NumberDisplay.slider,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed to use for generation.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "camera_fixed",
                     default=False,
                     tooltip="Specifies whether to fix the camera. The platform appends an instruction "
                             "to fix the camera to your prompt, but does not guarantee the actual effect.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "watermark",
                     default=True,
                     tooltip="Whether to add an \"AI generated\" watermark to the video.",
@@ -982,12 +982,12 @@ class ByteDanceFirstLastFrameNode(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -1005,7 +1005,7 @@ class ByteDanceFirstLastFrameNode(comfy_io.ComfyNode):
         seed: int,
         camera_fixed: bool,
         watermark: bool,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=True, min_length=1)
         raise_if_text_params(prompt, ["resolution", "ratio", "duration", "seed", "camerafixed", "watermark"])
         for i in (first_frame, last_frame):
@@ -1050,62 +1050,62 @@ class ByteDanceFirstLastFrameNode(comfy_io.ComfyNode):
         )
 
 
-class ByteDanceImageReferenceNode(comfy_io.ComfyNode):
+class ByteDanceImageReferenceNode(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="ByteDanceImageReferenceNode",
             display_name="ByteDance Reference Images to Video",
             category="api node/video/ByteDance",
             description="Generate video using prompt and reference images.",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=[Image2VideoModelName.seedance_1_lite.value],
                     default=Image2VideoModelName.seedance_1_lite.value,
                     tooltip="Model name",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     tooltip="The text prompt used to generate the video.",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "images",
                     tooltip="One to four images.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
                     options=["480p", "720p"],
                     tooltip="The resolution of the output video.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
                     options=["adaptive", "16:9", "4:3", "1:1", "3:4", "9:16", "21:9"],
                     tooltip="The aspect ratio of the output video.",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "duration",
                     default=5,
                     min=3,
                     max=12,
                     step=1,
                     tooltip="The duration of the output video in seconds.",
-                    display_mode=comfy_io.NumberDisplay.slider,
+                    display_mode=IO.NumberDisplay.slider,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed to use for generation.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "watermark",
                     default=True,
                     tooltip="Whether to add an \"AI generated\" watermark to the video.",
@@ -1113,12 +1113,12 @@ class ByteDanceImageReferenceNode(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -1134,7 +1134,7 @@ class ByteDanceImageReferenceNode(comfy_io.ComfyNode):
         duration: int,
         seed: int,
         watermark: bool,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=True, min_length=1)
         raise_if_text_params(prompt, ["resolution", "ratio", "duration", "seed", "watermark"])
         for image in images:
@@ -1180,7 +1180,7 @@ async def process_video_task(
     auth_kwargs: dict,
     node_id: str,
     estimated_duration: Optional[int],
-) -> comfy_io.NodeOutput:
+) -> IO.NodeOutput:
     initial_response = await SynchronousOperation(
         endpoint=ApiEndpoint(
             path=BYTEPLUS_TASK_ENDPOINT,
@@ -1197,7 +1197,7 @@ async def process_video_task(
         estimated_duration=estimated_duration,
         node_id=node_id,
     )
-    return comfy_io.NodeOutput(await download_url_to_video_output(get_video_url_from_task_status(response)))
+    return IO.NodeOutput(await download_url_to_video_output(get_video_url_from_task_status(response)))
 
 
 def raise_if_text_params(prompt: str, text_params: list[str]) -> None:
@@ -1210,7 +1210,7 @@ def raise_if_text_params(prompt: str, text_params: list[str]) -> None:
 
 class ByteDanceExtension(ComfyExtension):
     @override
-    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
         return [
             ByteDanceImageNode,
             ByteDanceImageEditNode,

comfy_api_nodes/nodes_gemini.py

@@ -26,7 +26,7 @@ from comfy_api_nodes.apis import (
     GeminiPart,
     GeminiMimeType,
 )
-from comfy_api_nodes.apis.gemini_api import GeminiImageGenerationConfig, GeminiImageGenerateContentRequest
+from comfy_api_nodes.apis.gemini_api import GeminiImageGenerationConfig, GeminiImageGenerateContentRequest, GeminiImageConfig
 from comfy_api_nodes.apis.client import (
     ApiEndpoint,
     HttpMethod,
@@ -63,6 +63,7 @@ class GeminiImageModel(str, Enum):
     """
 
     gemini_2_5_flash_image_preview = "gemini-2.5-flash-image-preview"
+    gemini_2_5_flash_image = "gemini-2.5-flash-image"
 
 
 def get_gemini_endpoint(
@@ -538,7 +539,7 @@ class GeminiImage(ComfyNodeABC):
                     {
                         "tooltip": "The Gemini model to use for generating responses.",
                         "options": [model.value for model in GeminiImageModel],
-                        "default": GeminiImageModel.gemini_2_5_flash_image_preview.value,
+                        "default": GeminiImageModel.gemini_2_5_flash_image.value,
                     },
                 ),
                 "seed": (
@@ -579,6 +580,14 @@ class GeminiImage(ComfyNodeABC):
                 #         "tooltip": "How many images to generate",
                 #     },
                 # ),
+                "aspect_ratio": (
+                    IO.COMBO,
+                    {
+                        "tooltip": "Defaults to matching the output image size to that of your input image, or otherwise generates 1:1 squares.",
+                        "options": ["auto", "1:1", "2:3", "3:2", "3:4", "4:3", "4:5", "5:4", "9:16", "16:9", "21:9"],
+                        "default": "auto",
+                    },
+                ),
             },
             "hidden": {
                 "auth_token": "AUTH_TOKEN_COMFY_ORG",
@@ -600,15 +609,17 @@ class GeminiImage(ComfyNodeABC):
         images: Optional[IO.IMAGE] = None,
         files: Optional[list[GeminiPart]] = None,
         n=1,
+        aspect_ratio: str = "auto",
         unique_id: Optional[str] = None,
         **kwargs,
     ):
-        # Validate inputs
         validate_string(prompt, strip_whitespace=True, min_length=1)
-        # Create parts list with text prompt as the first part
         parts: list[GeminiPart] = [create_text_part(prompt)]
 
-        # Add other modal parts
+        if not aspect_ratio:
+            aspect_ratio = "auto"  # for backward compatability with old workflows; to-do remove this in December
+        image_config = GeminiImageConfig(aspectRatio=aspect_ratio)
+
         if images is not None:
             image_parts = create_image_parts(images)
             parts.extend(image_parts)
@@ -625,7 +636,8 @@ class GeminiImage(ComfyNodeABC):
                     ),
                 ],
                 generationConfig=GeminiImageGenerationConfig(
-                    responseModalities=["TEXT","IMAGE"]
+                    responseModalities=["TEXT","IMAGE"],
+                    imageConfig=None if aspect_ratio == "auto" else image_config,
                 )
             ),
             auth_kwargs=kwargs,

comfy_api_nodes/nodes_ideogram.py

@@ -1,6 +1,6 @@
 from io import BytesIO
 from typing_extensions import override
-from comfy_api.latest import ComfyExtension, io as comfy_io
+from comfy_api.latest import ComfyExtension, IO
 from PIL import Image
 import numpy as np
 import torch
@@ -246,76 +246,76 @@ def display_image_urls_on_node(image_urls, node_id):
             PromptServer.instance.send_progress_text(urls_text, node_id)
 
 
-class IdeogramV1(comfy_io.ComfyNode):
+class IdeogramV1(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="IdeogramV1",
             display_name="Ideogram V1",
             category="api node/image/Ideogram",
             description="Generates images using the Ideogram V1 model.",
             is_api_node=True,
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the image generation",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "turbo",
                     default=False,
                     tooltip="Whether to use turbo mode (faster generation, potentially lower quality)",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
                     options=list(V1_V2_RATIO_MAP.keys()),
                     default="1:1",
                     tooltip="The aspect ratio for image generation.",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "magic_prompt_option",
                     options=["AUTO", "ON", "OFF"],
                     default="AUTO",
                     tooltip="Determine if MagicPrompt should be used in generation",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
                     control_after_generate=True,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     optional=True,
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     multiline=True,
                     default="",
                     tooltip="Description of what to exclude from the image",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "num_images",
                     default=1,
                     min=1,
                     max=8,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     optional=True,
                 ),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
         )
 
@@ -372,39 +372,39 @@ class IdeogramV1(comfy_io.ComfyNode):
             raise Exception("No image URLs were generated in the response")
 
         display_image_urls_on_node(image_urls, cls.hidden.unique_id)
-        return comfy_io.NodeOutput(await download_and_process_images(image_urls))
+        return IO.NodeOutput(await download_and_process_images(image_urls))
 
 
-class IdeogramV2(comfy_io.ComfyNode):
+class IdeogramV2(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="IdeogramV2",
             display_name="Ideogram V2",
             category="api node/image/Ideogram",
             description="Generates images using the Ideogram V2 model.",
             is_api_node=True,
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the image generation",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "turbo",
                     default=False,
                     tooltip="Whether to use turbo mode (faster generation, potentially lower quality)",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
                     options=list(V1_V2_RATIO_MAP.keys()),
                     default="1:1",
                     tooltip="The aspect ratio for image generation. Ignored if resolution is not set to AUTO.",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
                     options=list(V1_V1_RES_MAP.keys()),
                     default="Auto",
@@ -412,44 +412,44 @@ class IdeogramV2(comfy_io.ComfyNode):
                             "If not set to AUTO, this overrides the aspect_ratio setting.",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "magic_prompt_option",
                     options=["AUTO", "ON", "OFF"],
                     default="AUTO",
                     tooltip="Determine if MagicPrompt should be used in generation",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
                     control_after_generate=True,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "style_type",
                     options=["AUTO", "GENERAL", "REALISTIC", "DESIGN", "RENDER_3D", "ANIME"],
                     default="NONE",
                     tooltip="Style type for generation (V2 only)",
                     optional=True,
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     multiline=True,
                     default="",
                     tooltip="Description of what to exclude from the image",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "num_images",
                     default=1,
                     min=1,
                     max=8,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     optional=True,
                 ),
                 #"color_palette": (
@@ -462,12 +462,12 @@ class IdeogramV2(comfy_io.ComfyNode):
                 #),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
         )
 
@@ -541,14 +541,14 @@ class IdeogramV2(comfy_io.ComfyNode):
             raise Exception("No image URLs were generated in the response")
 
         display_image_urls_on_node(image_urls, cls.hidden.unique_id)
-        return comfy_io.NodeOutput(await download_and_process_images(image_urls))
+        return IO.NodeOutput(await download_and_process_images(image_urls))
 
 
-class IdeogramV3(comfy_io.ComfyNode):
+class IdeogramV3(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="IdeogramV3",
             display_name="Ideogram V3",
             category="api node/image/Ideogram",
@@ -556,30 +556,30 @@ class IdeogramV3(comfy_io.ComfyNode):
                         "Supports both regular image generation from text prompts and image editing with mask.",
             is_api_node=True,
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the image generation or editing",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                     tooltip="Optional reference image for image editing.",
                     optional=True,
                 ),
-                comfy_io.Mask.Input(
+                IO.Mask.Input(
                     "mask",
                     tooltip="Optional mask for inpainting (white areas will be replaced)",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
                     options=list(V3_RATIO_MAP.keys()),
                     default="1:1",
                     tooltip="The aspect ratio for image generation. Ignored if resolution is not set to Auto.",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
                     options=V3_RESOLUTIONS,
                     default="Auto",
@@ -587,57 +587,57 @@ class IdeogramV3(comfy_io.ComfyNode):
                             "If not set to Auto, this overrides the aspect_ratio setting.",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "magic_prompt_option",
                     options=["AUTO", "ON", "OFF"],
                     default="AUTO",
                     tooltip="Determine if MagicPrompt should be used in generation",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
                     control_after_generate=True,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "num_images",
                     default=1,
                     min=1,
                     max=8,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "rendering_speed",
                     options=["DEFAULT", "TURBO", "QUALITY"],
                     default="DEFAULT",
                     tooltip="Controls the trade-off between generation speed and quality",
                     optional=True,
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "character_image",
                     tooltip="Image to use as character reference.",
                     optional=True,
                 ),
-                comfy_io.Mask.Input(
+                IO.Mask.Input(
                     "character_mask",
                     tooltip="Optional mask for character reference image.",
                     optional=True,
                 ),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
         )
 
@@ -826,12 +826,12 @@ class IdeogramV3(comfy_io.ComfyNode):
             raise Exception("No image URLs were generated in the response")
 
         display_image_urls_on_node(image_urls, cls.hidden.unique_id)
-        return comfy_io.NodeOutput(await download_and_process_images(image_urls))
+        return IO.NodeOutput(await download_and_process_images(image_urls))
 
 
 class IdeogramExtension(ComfyExtension):
     @override
-    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
         return [
             IdeogramV1,
             IdeogramV2,

comfy_api_nodes/nodes_luma.py

@@ -2,7 +2,7 @@ from __future__ import annotations
 from inspect import cleandoc
 from typing import Optional
 from typing_extensions import override
-from comfy_api.latest import ComfyExtension, io as comfy_io
+from comfy_api.latest import ComfyExtension, IO
 from comfy_api.input_impl.video_types import VideoFromFile
 from comfy_api_nodes.apis.luma_api import (
     LumaImageModel,
@@ -52,24 +52,24 @@ def image_result_url_extractor(response: LumaGeneration):
 def video_result_url_extractor(response: LumaGeneration):
     return response.assets.video if hasattr(response, "assets") and hasattr(response.assets, "video") else None
 
-class LumaReferenceNode(comfy_io.ComfyNode):
+class LumaReferenceNode(IO.ComfyNode):
     """
     Holds an image and weight for use with Luma Generate Image node.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="LumaReferenceNode",
             display_name="Luma Reference",
             category="api node/image/Luma",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                     tooltip="Image to use as reference.",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "weight",
                     default=1.0,
                     min=0.0,
@@ -77,71 +77,71 @@ class LumaReferenceNode(comfy_io.ComfyNode):
                     step=0.01,
                     tooltip="Weight of image reference.",
                 ),
-                comfy_io.Custom(LumaIO.LUMA_REF).Input(
+                IO.Custom(LumaIO.LUMA_REF).Input(
                     "luma_ref",
                     optional=True,
                 ),
             ],
-            outputs=[comfy_io.Custom(LumaIO.LUMA_REF).Output(display_name="luma_ref")],
+            outputs=[IO.Custom(LumaIO.LUMA_REF).Output(display_name="luma_ref")],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
         )
 
     @classmethod
     def execute(
         cls, image: torch.Tensor, weight: float, luma_ref: LumaReferenceChain = None
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         if luma_ref is not None:
             luma_ref = luma_ref.clone()
         else:
             luma_ref = LumaReferenceChain()
         luma_ref.add(LumaReference(image=image, weight=round(weight, 2)))
-        return comfy_io.NodeOutput(luma_ref)
+        return IO.NodeOutput(luma_ref)
 
 
-class LumaConceptsNode(comfy_io.ComfyNode):
+class LumaConceptsNode(IO.ComfyNode):
     """
     Holds one or more Camera Concepts for use with Luma Text to Video and Luma Image to Video nodes.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="LumaConceptsNode",
             display_name="Luma Concepts",
             category="api node/video/Luma",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "concept1",
                     options=get_luma_concepts(include_none=True),
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "concept2",
                     options=get_luma_concepts(include_none=True),
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "concept3",
                     options=get_luma_concepts(include_none=True),
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "concept4",
                     options=get_luma_concepts(include_none=True),
                 ),
-                comfy_io.Custom(LumaIO.LUMA_CONCEPTS).Input(
+                IO.Custom(LumaIO.LUMA_CONCEPTS).Input(
                     "luma_concepts",
                     tooltip="Optional Camera Concepts to add to the ones chosen here.",
                     optional=True,
                 ),
             ],
-            outputs=[comfy_io.Custom(LumaIO.LUMA_CONCEPTS).Output(display_name="luma_concepts")],
+            outputs=[IO.Custom(LumaIO.LUMA_CONCEPTS).Output(display_name="luma_concepts")],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
         )
 
@@ -153,42 +153,42 @@ class LumaConceptsNode(comfy_io.ComfyNode):
         concept3: str,
         concept4: str,
         luma_concepts: LumaConceptChain = None,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         chain = LumaConceptChain(str_list=[concept1, concept2, concept3, concept4])
         if luma_concepts is not None:
             chain = luma_concepts.clone_and_merge(chain)
-        return comfy_io.NodeOutput(chain)
+        return IO.NodeOutput(chain)
 
 
-class LumaImageGenerationNode(comfy_io.ComfyNode):
+class LumaImageGenerationNode(IO.ComfyNode):
     """
     Generates images synchronously based on prompt and aspect ratio.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="LumaImageNode",
             display_name="Luma Text to Image",
             category="api node/image/Luma",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the image generation",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
-                    options=[model.value for model in LumaImageModel],
+                    options=LumaImageModel,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
-                    options=[ratio.value for ratio in LumaAspectRatio],
+                    options=LumaAspectRatio,
                     default=LumaAspectRatio.ratio_16_9,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -196,7 +196,7 @@ class LumaImageGenerationNode(comfy_io.ComfyNode):
                     control_after_generate=True,
                     tooltip="Seed to determine if node should re-run; actual results are nondeterministic regardless of seed.",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "style_image_weight",
                     default=1.0,
                     min=0.0,
@@ -204,27 +204,27 @@ class LumaImageGenerationNode(comfy_io.ComfyNode):
                     step=0.01,
                     tooltip="Weight of style image. Ignored if no style_image provided.",
                 ),
-                comfy_io.Custom(LumaIO.LUMA_REF).Input(
+                IO.Custom(LumaIO.LUMA_REF).Input(
                     "image_luma_ref",
                     tooltip="Luma Reference node connection to influence generation with input images; up to 4 images can be considered.",
                     optional=True,
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "style_image",
                     tooltip="Style reference image; only 1 image will be used.",
                     optional=True,
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "character_image",
                     tooltip="Character reference images; can be a batch of multiple, up to 4 images can be considered.",
                     optional=True,
                 ),
             ],
-            outputs=[comfy_io.Image.Output()],
+            outputs=[IO.Image.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -240,7 +240,7 @@ class LumaImageGenerationNode(comfy_io.ComfyNode):
         image_luma_ref: LumaReferenceChain = None,
         style_image: torch.Tensor = None,
         character_image: torch.Tensor = None,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=True, min_length=3)
         auth_kwargs = {
             "auth_token": cls.hidden.auth_token_comfy_org,
@@ -306,7 +306,7 @@ class LumaImageGenerationNode(comfy_io.ComfyNode):
         async with aiohttp.ClientSession() as session:
             async with session.get(response_poll.assets.image) as img_response:
                 img = process_image_response(await img_response.content.read())
-        return comfy_io.NodeOutput(img)
+        return IO.NodeOutput(img)
 
     @classmethod
     async def _convert_luma_refs(
@@ -334,29 +334,29 @@ class LumaImageGenerationNode(comfy_io.ComfyNode):
         return await cls._convert_luma_refs(chain, max_refs=1, auth_kwargs=auth_kwargs)
 
 
-class LumaImageModifyNode(comfy_io.ComfyNode):
+class LumaImageModifyNode(IO.ComfyNode):
     """
     Modifies images synchronously based on prompt and aspect ratio.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="LumaImageModifyNode",
             display_name="Luma Image to Image",
             category="api node/image/Luma",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the image generation",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "image_weight",
                     default=0.1,
                     min=0.0,
@@ -364,11 +364,11 @@ class LumaImageModifyNode(comfy_io.ComfyNode):
                     step=0.01,
                     tooltip="Weight of the image; the closer to 1.0, the less the image will be modified.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
-                    options=[model.value for model in LumaImageModel],
+                    options=LumaImageModel,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -377,11 +377,11 @@ class LumaImageModifyNode(comfy_io.ComfyNode):
                     tooltip="Seed to determine if node should re-run; actual results are nondeterministic regardless of seed.",
                 ),
             ],
-            outputs=[comfy_io.Image.Output()],
+            outputs=[IO.Image.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -394,7 +394,7 @@ class LumaImageModifyNode(comfy_io.ComfyNode):
         image: torch.Tensor,
         image_weight: float,
         seed,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         auth_kwargs = {
             "auth_token": cls.hidden.auth_token_comfy_org,
             "comfy_api_key": cls.hidden.api_key_comfy_org,
@@ -442,51 +442,51 @@ class LumaImageModifyNode(comfy_io.ComfyNode):
         async with aiohttp.ClientSession() as session:
             async with session.get(response_poll.assets.image) as img_response:
                 img = process_image_response(await img_response.content.read())
-        return comfy_io.NodeOutput(img)
+        return IO.NodeOutput(img)
 
 
-class LumaTextToVideoGenerationNode(comfy_io.ComfyNode):
+class LumaTextToVideoGenerationNode(IO.ComfyNode):
     """
     Generates videos synchronously based on prompt and output_size.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="LumaVideoNode",
             display_name="Luma Text to Video",
             category="api node/video/Luma",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the video generation",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
-                    options=[model.value for model in LumaVideoModel],
+                    options=LumaVideoModel,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
-                    options=[ratio.value for ratio in LumaAspectRatio],
+                    options=LumaAspectRatio,
                     default=LumaAspectRatio.ratio_16_9,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
-                    options=[resolution.value for resolution in LumaVideoOutputResolution],
+                    options=LumaVideoOutputResolution,
                     default=LumaVideoOutputResolution.res_540p,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "duration",
-                    options=[dur.value for dur in LumaVideoModelOutputDuration],
+                    options=LumaVideoModelOutputDuration,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "loop",
                     default=False,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -494,17 +494,17 @@ class LumaTextToVideoGenerationNode(comfy_io.ComfyNode):
                     control_after_generate=True,
                     tooltip="Seed to determine if node should re-run; actual results are nondeterministic regardless of seed.",
                 ),
-                comfy_io.Custom(LumaIO.LUMA_CONCEPTS).Input(
+                IO.Custom(LumaIO.LUMA_CONCEPTS).Input(
                     "luma_concepts",
                     tooltip="Optional Camera Concepts to dictate camera motion via the Luma Concepts node.",
                     optional=True,
                 )
             ],
-            outputs=[comfy_io.Video.Output()],
+            outputs=[IO.Video.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -520,7 +520,7 @@ class LumaTextToVideoGenerationNode(comfy_io.ComfyNode):
         loop: bool,
         seed,
         luma_concepts: LumaConceptChain = None,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=False, min_length=3)
         duration = duration if model != LumaVideoModel.ray_1_6 else None
         resolution = resolution if model != LumaVideoModel.ray_1_6 else None
@@ -571,51 +571,51 @@ class LumaTextToVideoGenerationNode(comfy_io.ComfyNode):
 
         async with aiohttp.ClientSession() as session:
             async with session.get(response_poll.assets.video) as vid_response:
-                return comfy_io.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read())))
+                return IO.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read())))
 
 
-class LumaImageToVideoGenerationNode(comfy_io.ComfyNode):
+class LumaImageToVideoGenerationNode(IO.ComfyNode):
     """
     Generates videos synchronously based on prompt, input images, and output_size.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="LumaImageToVideoNode",
             display_name="Luma Image to Video",
             category="api node/video/Luma",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the video generation",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
-                    options=[model.value for model in LumaVideoModel],
+                    options=LumaVideoModel,
                 ),
-                # comfy_io.Combo.Input(
+                # IO.Combo.Input(
                 #     "aspect_ratio",
                 #     options=[ratio.value for ratio in LumaAspectRatio],
                 #     default=LumaAspectRatio.ratio_16_9,
                 # ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
-                    options=[resolution.value for resolution in LumaVideoOutputResolution],
+                    options=LumaVideoOutputResolution,
                     default=LumaVideoOutputResolution.res_540p,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "duration",
                     options=[dur.value for dur in LumaVideoModelOutputDuration],
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "loop",
                     default=False,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -623,27 +623,27 @@ class LumaImageToVideoGenerationNode(comfy_io.ComfyNode):
                     control_after_generate=True,
                     tooltip="Seed to determine if node should re-run; actual results are nondeterministic regardless of seed.",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "first_image",
                     tooltip="First frame of generated video.",
                     optional=True,
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "last_image",
                     tooltip="Last frame of generated video.",
                     optional=True,
                 ),
-                comfy_io.Custom(LumaIO.LUMA_CONCEPTS).Input(
+                IO.Custom(LumaIO.LUMA_CONCEPTS).Input(
                     "luma_concepts",
                     tooltip="Optional Camera Concepts to dictate camera motion via the Luma Concepts node.",
                     optional=True,
                 )
             ],
-            outputs=[comfy_io.Video.Output()],
+            outputs=[IO.Video.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -660,7 +660,7 @@ class LumaImageToVideoGenerationNode(comfy_io.ComfyNode):
         first_image: torch.Tensor = None,
         last_image: torch.Tensor = None,
         luma_concepts: LumaConceptChain = None,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         if first_image is None and last_image is None:
             raise Exception(
                 "At least one of first_image and last_image requires an input."
@@ -716,7 +716,7 @@ class LumaImageToVideoGenerationNode(comfy_io.ComfyNode):
 
         async with aiohttp.ClientSession() as session:
             async with session.get(response_poll.assets.video) as vid_response:
-                return comfy_io.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read())))
+                return IO.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read())))
 
     @classmethod
     async def _convert_to_keyframes(
@@ -744,7 +744,7 @@ class LumaImageToVideoGenerationNode(comfy_io.ComfyNode):
 
 class LumaExtension(ComfyExtension):
     @override
-    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
         return [
             LumaImageGenerationNode,
             LumaImageModifyNode,

comfy_api_nodes/nodes_minimax.py

@@ -4,7 +4,7 @@ import logging
 import torch
 
 from typing_extensions import override
-from comfy_api.latest import ComfyExtension, io as comfy_io
+from comfy_api.latest import ComfyExtension, IO
 from comfy_api.input_impl.video_types import VideoFromFile
 from comfy_api_nodes.apis import (
     MinimaxVideoGenerationRequest,
@@ -43,7 +43,7 @@ async def _generate_mm_video(
     image: Optional[torch.Tensor] = None,   # used for ImageToVideo
     subject: Optional[torch.Tensor] = None, # used for SubjectToVideo
     average_duration: Optional[int] = None,
-) -> comfy_io.NodeOutput:
+) -> IO.NodeOutput:
     if image is None:
         validate_string(prompt_text, field_name="prompt_text")
     # upload image, if passed in
@@ -133,35 +133,35 @@ async def _generate_mm_video(
         error_msg = f"Failed to download video from {file_url}"
         logging.error(error_msg)
         raise Exception(error_msg)
-    return comfy_io.NodeOutput(VideoFromFile(video_io))
+    return IO.NodeOutput(VideoFromFile(video_io))
 
 
-class MinimaxTextToVideoNode(comfy_io.ComfyNode):
+class MinimaxTextToVideoNode(IO.ComfyNode):
     """
     Generates videos synchronously based on a prompt, and optional parameters using MiniMax's API.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="MinimaxTextToVideoNode",
             display_name="MiniMax Text to Video",
             category="api node/video/MiniMax",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt_text",
                     multiline=True,
                     default="",
                     tooltip="Text prompt to guide the video generation",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=["T2V-01", "T2V-01-Director"],
                     default="T2V-01",
                     tooltip="Model to use for video generation",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -172,11 +172,11 @@ class MinimaxTextToVideoNode(comfy_io.ComfyNode):
                     optional=True,
                 ),
             ],
-            outputs=[comfy_io.Video.Output()],
+            outputs=[IO.Video.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -187,7 +187,7 @@ class MinimaxTextToVideoNode(comfy_io.ComfyNode):
         prompt_text: str,
         model: str = "T2V-01",
         seed: int = 0,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         return await _generate_mm_video(
             auth={
                 "auth_token": cls.hidden.auth_token_comfy_org,
@@ -203,36 +203,36 @@ class MinimaxTextToVideoNode(comfy_io.ComfyNode):
         )
 
 
-class MinimaxImageToVideoNode(comfy_io.ComfyNode):
+class MinimaxImageToVideoNode(IO.ComfyNode):
     """
     Generates videos synchronously based on an image and prompt, and optional parameters using MiniMax's API.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="MinimaxImageToVideoNode",
             display_name="MiniMax Image to Video",
             category="api node/video/MiniMax",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                     tooltip="Image to use as first frame of video generation",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt_text",
                     multiline=True,
                     default="",
                     tooltip="Text prompt to guide the video generation",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=["I2V-01-Director", "I2V-01", "I2V-01-live"],
                     default="I2V-01",
                     tooltip="Model to use for video generation",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -243,11 +243,11 @@ class MinimaxImageToVideoNode(comfy_io.ComfyNode):
                     optional=True,
                 ),
             ],
-            outputs=[comfy_io.Video.Output()],
+            outputs=[IO.Video.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -259,7 +259,7 @@ class MinimaxImageToVideoNode(comfy_io.ComfyNode):
         prompt_text: str,
         model: str = "I2V-01",
         seed: int = 0,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         return await _generate_mm_video(
             auth={
                 "auth_token": cls.hidden.auth_token_comfy_org,
@@ -275,36 +275,36 @@ class MinimaxImageToVideoNode(comfy_io.ComfyNode):
         )
 
 
-class MinimaxSubjectToVideoNode(comfy_io.ComfyNode):
+class MinimaxSubjectToVideoNode(IO.ComfyNode):
     """
     Generates videos synchronously based on an image and prompt, and optional parameters using MiniMax's API.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="MinimaxSubjectToVideoNode",
             display_name="MiniMax Subject to Video",
             category="api node/video/MiniMax",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "subject",
                     tooltip="Image of subject to reference for video generation",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt_text",
                     multiline=True,
                     default="",
                     tooltip="Text prompt to guide the video generation",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=["S2V-01"],
                     default="S2V-01",
                     tooltip="Model to use for video generation",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -315,11 +315,11 @@ class MinimaxSubjectToVideoNode(comfy_io.ComfyNode):
                     optional=True,
                 ),
             ],
-            outputs=[comfy_io.Video.Output()],
+            outputs=[IO.Video.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -331,7 +331,7 @@ class MinimaxSubjectToVideoNode(comfy_io.ComfyNode):
         prompt_text: str,
         model: str = "S2V-01",
         seed: int = 0,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         return await _generate_mm_video(
             auth={
                 "auth_token": cls.hidden.auth_token_comfy_org,
@@ -347,24 +347,24 @@ class MinimaxSubjectToVideoNode(comfy_io.ComfyNode):
         )
 
 
-class MinimaxHailuoVideoNode(comfy_io.ComfyNode):
+class MinimaxHailuoVideoNode(IO.ComfyNode):
     """Generates videos from prompt, with optional start frame using the new MiniMax Hailuo-02 model."""
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="MinimaxHailuoVideoNode",
             display_name="MiniMax Hailuo Video",
             category="api node/video/MiniMax",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt_text",
                     multiline=True,
                     default="",
                     tooltip="Text prompt to guide the video generation.",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -374,25 +374,25 @@ class MinimaxHailuoVideoNode(comfy_io.ComfyNode):
                     tooltip="The random seed used for creating the noise.",
                     optional=True,
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "first_frame_image",
                     tooltip="Optional image to use as the first frame to generate a video.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "prompt_optimizer",
                     default=True,
                     tooltip="Optimize prompt to improve generation quality when needed.",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "duration",
                     options=[6, 10],
                     default=6,
                     tooltip="The length of the output video in seconds.",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
                     options=["768P", "1080P"],
                     default="768P",
@@ -400,11 +400,11 @@ class MinimaxHailuoVideoNode(comfy_io.ComfyNode):
                     optional=True,
                 ),
             ],
-            outputs=[comfy_io.Video.Output()],
+            outputs=[IO.Video.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -419,7 +419,7 @@ class MinimaxHailuoVideoNode(comfy_io.ComfyNode):
         duration: int = 6,
         resolution: str = "768P",
         model: str = "MiniMax-Hailuo-02",
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         auth = {
             "auth_token": cls.hidden.auth_token_comfy_org,
             "comfy_api_key": cls.hidden.api_key_comfy_org,
@@ -500,7 +500,7 @@ class MinimaxHailuoVideoNode(comfy_io.ComfyNode):
             raise Exception(
                 f"No video was found in the response. Full response: {file_result.model_dump()}"
             )
-        logging.info(f"Generated video URL: {file_url}")
+        logging.info("Generated video URL: %s", file_url)
         if cls.hidden.unique_id:
             if hasattr(file_result.file, "backup_download_url"):
                 message = f"Result URL: {file_url}\nBackup URL: {file_result.file.backup_download_url}"
@@ -513,12 +513,12 @@ class MinimaxHailuoVideoNode(comfy_io.ComfyNode):
             error_msg = f"Failed to download video from {file_url}"
             logging.error(error_msg)
             raise Exception(error_msg)
-        return comfy_io.NodeOutput(VideoFromFile(video_io))
+        return IO.NodeOutput(VideoFromFile(video_io))
 
 
 class MinimaxExtension(ComfyExtension):
     @override
-    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
         return [
             MinimaxTextToVideoNode,
             MinimaxImageToVideoNode,

comfy_api_nodes/nodes_moonvalley.py

@@ -22,10 +22,11 @@ from comfy_api_nodes.apinode_utils import (
     download_url_to_video_output,
     upload_images_to_comfyapi,
     upload_video_to_comfyapi,
+    validate_container_format_is_mp4,
 )
 
 from comfy_api.input import VideoInput
-from comfy_api.latest import ComfyExtension, InputImpl, io as comfy_io
+from comfy_api.latest import ComfyExtension, InputImpl, IO
 import av
 import io
 
@@ -144,7 +145,7 @@ def validate_video_to_video_input(video: VideoInput) -> VideoInput:
     """
     width, height = _get_video_dimensions(video)
     _validate_video_dimensions(width, height)
-    _validate_container_format(video)
+    validate_container_format_is_mp4(video)
 
     return _validate_and_trim_duration(video)
 
@@ -177,15 +178,6 @@ def _validate_video_dimensions(width: int, height: int) -> None:
         )
 
 
-def _validate_container_format(video: VideoInput) -> None:
-    """Validates video container format is MP4."""
-    container_format = video.get_container_format()
-    if container_format not in ["mp4", "mov,mp4,m4a,3gp,3g2,mj2"]:
-        raise ValueError(
-            f"Only MP4 container format supported. Got: {container_format}"
-        )
-
-
 def _validate_and_trim_duration(video: VideoInput) -> VideoInput:
     """Validates video duration and trims to 5 seconds if needed."""
     duration = video.get_duration()
@@ -237,7 +229,7 @@ def trim_video(video: VideoInput, duration_sec: float) -> VideoInput:
         audio_stream = None
 
         for stream in input_container.streams:
-            logging.info(f"Found stream: type={stream.type}, class={type(stream)}")
+            logging.info("Found stream: type=%s, class=%s", stream.type, type(stream))
             if isinstance(stream, av.VideoStream):
                 # Create output video stream with same parameters
                 video_stream = output_container.add_stream(
@@ -247,7 +239,7 @@ def trim_video(video: VideoInput, duration_sec: float) -> VideoInput:
                 video_stream.height = stream.height
                 video_stream.pix_fmt = "yuv420p"
                 logging.info(
-                    f"Added video stream: {stream.width}x{stream.height} @ {stream.average_rate}fps"
+                    "Added video stream: %sx%s @ %sfps", stream.width, stream.height, stream.average_rate
                 )
             elif isinstance(stream, av.AudioStream):
                 # Create output audio stream with same parameters
@@ -256,9 +248,7 @@ def trim_video(video: VideoInput, duration_sec: float) -> VideoInput:
                 )
                 audio_stream.sample_rate = stream.sample_rate
                 audio_stream.layout = stream.layout
-                logging.info(
-                    f"Added audio stream: {stream.sample_rate}Hz, {stream.channels} channels"
-                )
+                logging.info("Added audio stream: %sHz, %s channels", stream.sample_rate, stream.channels)
 
         # Calculate target frame count that's divisible by 16
         fps = input_container.streams.video[0].average_rate
@@ -288,9 +278,7 @@ def trim_video(video: VideoInput, duration_sec: float) -> VideoInput:
             for packet in video_stream.encode():
                 output_container.mux(packet)
 
-            logging.info(
-                f"Encoded {frame_count} video frames (target: {target_frames})"
-            )
+            logging.info("Encoded %s video frames (target: %s)", frame_count, target_frames)
 
         # Decode and re-encode audio frames
         if audio_stream:
@@ -308,7 +296,7 @@ def trim_video(video: VideoInput, duration_sec: float) -> VideoInput:
             for packet in audio_stream.encode():
                 output_container.mux(packet)
 
-            logging.info(f"Encoded {audio_frame_count} audio frames")
+            logging.info("Encoded %s audio frames", audio_frame_count)
 
         # Close containers
         output_container.close()
@@ -335,7 +323,7 @@ def parse_width_height_from_res(resolution: str):
         "1:1 (1152 x 1152)": {"width": 1152, "height": 1152},
         "4:3 (1536 x 1152)": {"width": 1536, "height": 1152},
         "3:4 (1152 x 1536)": {"width": 1152, "height": 1536},
-        "21:9 (2560 x 1080)": {"width": 2560, "height": 1080},
+        # "21:9 (2560 x 1080)": {"width": 2560, "height": 1080},
     }
     return res_map.get(resolution, {"width": 1920, "height": 1080})
 
@@ -366,36 +354,36 @@ async def get_response(
     )
 
 
-class MoonvalleyImg2VideoNode(comfy_io.ComfyNode):
+class MoonvalleyImg2VideoNode(IO.ComfyNode):
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="MoonvalleyImg2VideoNode",
             display_name="Moonvalley Marey Image to Video",
             category="api node/video/Moonvalley Marey",
             description="Moonvalley Marey Image to Video Node",
             inputs=[
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                     tooltip="The reference image used to generate the video",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     multiline=True,
                     default="<synthetic> <scene cut> gopro, bright, contrast, static, overexposed, vignette, "
-                            "artifacts, still, noise, texture, scanlines, videogame, 360 camera, VR, transition, "
-                            "flare, saturation, distorted, warped, wide angle, saturated, vibrant, glowing, "
-                            "cross dissolve, cheesy, ugly hands, mutated hands, mutant, disfigured, extra fingers, "
-                            "blown out, horrible, blurry, worst quality, bad, dissolve, melt, fade in, fade out, "
-                            "wobbly, weird, low quality, plastic, stock footage, video camera, boring",
+                    "artifacts, still, noise, texture, scanlines, videogame, 360 camera, VR, transition, "
+                    "flare, saturation, distorted, warped, wide angle, saturated, vibrant, glowing, "
+                    "cross dissolve, cheesy, ugly hands, mutated hands, mutant, disfigured, extra fingers, "
+                    "blown out, horrible, blurry, worst quality, bad, dissolve, melt, fade in, fade out, "
+                    "wobbly, weird, low quality, plastic, stock footage, video camera, boring",
                     tooltip="Negative prompt text",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
                     options=[
                         "16:9 (1920 x 1080)",
@@ -403,42 +391,43 @@ class MoonvalleyImg2VideoNode(comfy_io.ComfyNode):
                         "1:1 (1152 x 1152)",
                         "4:3 (1536 x 1152)",
                         "3:4 (1152 x 1536)",
-                        "21:9 (2560 x 1080)",
+                        # "21:9 (2560 x 1080)",
                     ],
                     default="16:9 (1920 x 1080)",
                     tooltip="Resolution of the output video",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "prompt_adherence",
-                    default=10.0,
+                    default=4.5,
                     min=1.0,
                     max=20.0,
                     step=1.0,
                     tooltip="Guidance scale for generation control",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=9,
                     min=0,
                     max=4294967295,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Random seed value",
+                    control_after_generate=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "steps",
-                    default=100,
+                    default=33,
                     min=1,
                     max=100,
                     step=1,
                     tooltip="Number of denoising steps",
                 ),
             ],
-            outputs=[comfy_io.Video.Output()],
+            outputs=[IO.Video.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -453,7 +442,7 @@ class MoonvalleyImg2VideoNode(comfy_io.ComfyNode):
         prompt_adherence: float,
         seed: int,
         steps: int,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_image_dimensions(image, min_width=300, min_height=300, max_height=MAX_HEIGHT, max_width=MAX_WIDTH)
         validate_prompts(prompt, negative_prompt, MOONVALLEY_MAREY_MAX_PROMPT_LENGTH)
         width_height = parse_width_height_from_res(resolution)
@@ -468,12 +457,11 @@ class MoonvalleyImg2VideoNode(comfy_io.ComfyNode):
             steps=steps,
             seed=seed,
             guidance_scale=prompt_adherence,
-            num_frames=128,
             width=width_height["width"],
             height=width_height["height"],
             use_negative_prompts=True,
         )
-        """Upload image to comfy backend to have a URL available for further processing"""
+
         # Get MIME type from tensor - assuming PNG format for image tensors
         mime_type = "image/png"
 
@@ -504,57 +492,57 @@ class MoonvalleyImg2VideoNode(comfy_io.ComfyNode):
             task_id, auth_kwargs=auth, node_id=cls.hidden.unique_id
         )
         video = await download_url_to_video_output(final_response.output_url)
-        return comfy_io.NodeOutput(video)
+        return IO.NodeOutput(video)
 
 
-class MoonvalleyVideo2VideoNode(comfy_io.ComfyNode):
+class MoonvalleyVideo2VideoNode(IO.ComfyNode):
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="MoonvalleyVideo2VideoNode",
             display_name="Moonvalley Marey Video to Video",
             category="api node/video/Moonvalley Marey",
             description="",
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     tooltip="Describes the video to generate",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     multiline=True,
                     default="<synthetic> <scene cut> gopro, bright, contrast, static, overexposed, vignette, "
-                            "artifacts, still, noise, texture, scanlines, videogame, 360 camera, VR, transition, "
-                            "flare, saturation, distorted, warped, wide angle, saturated, vibrant, glowing, "
-                            "cross dissolve, cheesy, ugly hands, mutated hands, mutant, disfigured, extra fingers, "
-                            "blown out, horrible, blurry, worst quality, bad, dissolve, melt, fade in, fade out, "
-                            "wobbly, weird, low quality, plastic, stock footage, video camera, boring",
+                    "artifacts, still, noise, texture, scanlines, videogame, 360 camera, VR, transition, "
+                    "flare, saturation, distorted, warped, wide angle, saturated, vibrant, glowing, "
+                    "cross dissolve, cheesy, ugly hands, mutated hands, mutant, disfigured, extra fingers, "
+                    "blown out, horrible, blurry, worst quality, bad, dissolve, melt, fade in, fade out, "
+                    "wobbly, weird, low quality, plastic, stock footage, video camera, boring",
                     tooltip="Negative prompt text",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=9,
                     min=0,
                     max=4294967295,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Random seed value",
                     control_after_generate=False,
                 ),
-                comfy_io.Video.Input(
+                IO.Video.Input(
                     "video",
                     tooltip="The reference video used to generate the output video. Must be at least 5 seconds long. "
-                            "Videos longer than 5s will be automatically trimmed. Only MP4 format supported.",
+                    "Videos longer than 5s will be automatically trimmed. Only MP4 format supported.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "control_type",
                     options=["Motion Transfer", "Pose Transfer"],
                     default="Motion Transfer",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "motion_intensity",
                     default=100,
                     min=0,
@@ -563,12 +551,21 @@ class MoonvalleyVideo2VideoNode(comfy_io.ComfyNode):
                     tooltip="Only used if control_type is 'Motion Transfer'",
                     optional=True,
                 ),
+                IO.Int.Input(
+                    "steps",
+                    default=33,
+                    min=1,
+                    max=100,
+                    step=1,
+                    display_mode=IO.NumberDisplay.number,
+                    tooltip="Number of inference steps",
+                ),
             ],
-            outputs=[comfy_io.Video.Output()],
+            outputs=[IO.Video.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -582,7 +579,9 @@ class MoonvalleyVideo2VideoNode(comfy_io.ComfyNode):
         video: Optional[VideoInput] = None,
         control_type: str = "Motion Transfer",
         motion_intensity: Optional[int] = 100,
-    ) -> comfy_io.NodeOutput:
+        steps=33,
+        prompt_adherence=4.5,
+    ) -> IO.NodeOutput:
         auth = {
             "auth_token": cls.hidden.auth_token_comfy_org,
             "comfy_api_key": cls.hidden.api_key_comfy_org,
@@ -591,7 +590,6 @@ class MoonvalleyVideo2VideoNode(comfy_io.ComfyNode):
         validated_video = validate_video_to_video_input(video)
         video_url = await upload_video_to_comfyapi(validated_video, auth_kwargs=auth)
 
-        """Validate prompts and inference input"""
         validate_prompts(prompt, negative_prompt)
 
         # Only include motion_intensity for Motion Transfer
@@ -603,6 +601,8 @@ class MoonvalleyVideo2VideoNode(comfy_io.ComfyNode):
             negative_prompt=negative_prompt,
             seed=seed,
             control_params=control_params,
+            steps=steps,
+            guidance_scale=prompt_adherence,
         )
 
         control = parse_control_parameter(control_type)
@@ -633,35 +633,35 @@ class MoonvalleyVideo2VideoNode(comfy_io.ComfyNode):
         )
 
         video = await download_url_to_video_output(final_response.output_url)
-        return comfy_io.NodeOutput(video)
+        return IO.NodeOutput(video)
 
 
-class MoonvalleyTxt2VideoNode(comfy_io.ComfyNode):
+class MoonvalleyTxt2VideoNode(IO.ComfyNode):
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="MoonvalleyTxt2VideoNode",
             display_name="Moonvalley Marey Text to Video",
             category="api node/video/Moonvalley Marey",
             description="",
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     multiline=True,
                     default="<synthetic> <scene cut> gopro, bright, contrast, static, overexposed, vignette, "
-                            "artifacts, still, noise, texture, scanlines, videogame, 360 camera, VR, transition, "
-                            "flare, saturation, distorted, warped, wide angle, saturated, vibrant, glowing, "
-                            "cross dissolve, cheesy, ugly hands, mutated hands, mutant, disfigured, extra fingers, "
-                            "blown out, horrible, blurry, worst quality, bad, dissolve, melt, fade in, fade out, "
-                            "wobbly, weird, low quality, plastic, stock footage, video camera, boring",
+                    "artifacts, still, noise, texture, scanlines, videogame, 360 camera, VR, transition, "
+                    "flare, saturation, distorted, warped, wide angle, saturated, vibrant, glowing, "
+                    "cross dissolve, cheesy, ugly hands, mutated hands, mutant, disfigured, extra fingers, "
+                    "blown out, horrible, blurry, worst quality, bad, dissolve, melt, fade in, fade out, "
+                    "wobbly, weird, low quality, plastic, stock footage, video camera, boring",
                     tooltip="Negative prompt text",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
                     options=[
                         "16:9 (1920 x 1080)",
@@ -674,37 +674,38 @@ class MoonvalleyTxt2VideoNode(comfy_io.ComfyNode):
                     default="16:9 (1920 x 1080)",
                     tooltip="Resolution of the output video",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "prompt_adherence",
-                    default=10.0,
+                    default=4.0,
                     min=1.0,
                     max=20.0,
                     step=1.0,
                     tooltip="Guidance scale for generation control",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=9,
                     min=0,
                     max=4294967295,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
+                    control_after_generate=True,
                     tooltip="Random seed value",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "steps",
-                    default=100,
+                    default=33,
                     min=1,
                     max=100,
                     step=1,
                     tooltip="Inference steps",
                 ),
             ],
-            outputs=[comfy_io.Video.Output()],
+            outputs=[IO.Video.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -718,7 +719,7 @@ class MoonvalleyTxt2VideoNode(comfy_io.ComfyNode):
         prompt_adherence: float,
         seed: int,
         steps: int,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_prompts(prompt, negative_prompt, MOONVALLEY_MAREY_MAX_PROMPT_LENGTH)
         width_height = parse_width_height_from_res(resolution)
 
@@ -759,12 +760,12 @@ class MoonvalleyTxt2VideoNode(comfy_io.ComfyNode):
         )
 
         video = await download_url_to_video_output(final_response.output_url)
-        return comfy_io.NodeOutput(video)
+        return IO.NodeOutput(video)
 
 
 class MoonvalleyExtension(ComfyExtension):
     @override
-    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
         return [
             MoonvalleyImg2VideoNode,
             MoonvalleyTxt2VideoNode,

comfy_api_nodes/nodes_pixverse.py

@@ -29,7 +29,7 @@ from comfy_api_nodes.apinode_utils import (
     validate_string,
 )
 from comfy_api.input_impl import VideoFromFile
-from comfy_api.latest import ComfyExtension, io as comfy_io
+from comfy_api.latest import ComfyExtension, IO
 
 import torch
 import aiohttp
@@ -73,69 +73,69 @@ async def upload_image_to_pixverse(image: torch.Tensor, auth_kwargs=None):
     return response_upload.Resp.img_id
 
 
-class PixverseTemplateNode(comfy_io.ComfyNode):
+class PixverseTemplateNode(IO.ComfyNode):
     """
     Select template for PixVerse Video generation.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="PixverseTemplateNode",
             display_name="PixVerse Template",
             category="api node/video/PixVerse",
             inputs=[
-                comfy_io.Combo.Input("template", options=[list(pixverse_templates.keys())]),
+                IO.Combo.Input("template", options=list(pixverse_templates.keys())),
             ],
-            outputs=[comfy_io.Custom(PixverseIO.TEMPLATE).Output(display_name="pixverse_template")],
+            outputs=[IO.Custom(PixverseIO.TEMPLATE).Output(display_name="pixverse_template")],
         )
 
     @classmethod
-    def execute(cls, template: str) -> comfy_io.NodeOutput:
+    def execute(cls, template: str) -> IO.NodeOutput:
         template_id = pixverse_templates.get(template, None)
         if template_id is None:
             raise Exception(f"Template '{template}' is not recognized.")
         # just return the integer
-        return comfy_io.NodeOutput(template_id)
+        return IO.NodeOutput(template_id)
 
 
-class PixverseTextToVideoNode(comfy_io.ComfyNode):
+class PixverseTextToVideoNode(IO.ComfyNode):
     """
     Generates videos based on prompt and output_size.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="PixverseTextToVideoNode",
             display_name="PixVerse Text to Video",
             category="api node/video/PixVerse",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the video generation",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
-                    options=[ratio.value for ratio in PixverseAspectRatio],
+                    options=PixverseAspectRatio,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "quality",
-                    options=[resolution.value for resolution in PixverseQuality],
+                    options=PixverseQuality,
                     default=PixverseQuality.res_540p,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "duration_seconds",
-                    options=[dur.value for dur in PixverseDuration],
+                    options=PixverseDuration,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "motion_mode",
-                    options=[mode.value for mode in PixverseMotionMode],
+                    options=PixverseMotionMode,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -143,24 +143,24 @@ class PixverseTextToVideoNode(comfy_io.ComfyNode):
                     control_after_generate=True,
                     tooltip="Seed for video generation.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     default="",
-                    force_input=True,
+                    multiline=True,
                     tooltip="An optional text description of undesired elements on an image.",
                     optional=True,
                 ),
-                comfy_io.Custom(PixverseIO.TEMPLATE).Input(
+                IO.Custom(PixverseIO.TEMPLATE).Input(
                     "pixverse_template",
                     tooltip="An optional template to influence style of generation, created by the PixVerse Template node.",
                     optional=True,
                 ),
             ],
-            outputs=[comfy_io.Video.Output()],
+            outputs=[IO.Video.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -176,7 +176,7 @@ class PixverseTextToVideoNode(comfy_io.ComfyNode):
         seed,
         negative_prompt: str = None,
         pixverse_template: int = None,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=False)
         # 1080p is limited to 5 seconds duration
         # only normal motion_mode supported for 1080p or for non-5 second duration
@@ -237,43 +237,43 @@ class PixverseTextToVideoNode(comfy_io.ComfyNode):
 
         async with aiohttp.ClientSession() as session:
             async with session.get(response_poll.Resp.url) as vid_response:
-                return comfy_io.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read())))
+                return IO.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read())))
 
 
-class PixverseImageToVideoNode(comfy_io.ComfyNode):
+class PixverseImageToVideoNode(IO.ComfyNode):
     """
     Generates videos based on prompt and output_size.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="PixverseImageToVideoNode",
             display_name="PixVerse Image to Video",
             category="api node/video/PixVerse",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("image"),
-                comfy_io.String.Input(
+                IO.Image.Input("image"),
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the video generation",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "quality",
-                    options=[resolution.value for resolution in PixverseQuality],
+                    options=PixverseQuality,
                     default=PixverseQuality.res_540p,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "duration_seconds",
-                    options=[dur.value for dur in PixverseDuration],
+                    options=PixverseDuration,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "motion_mode",
-                    options=[mode.value for mode in PixverseMotionMode],
+                    options=PixverseMotionMode,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -281,24 +281,24 @@ class PixverseImageToVideoNode(comfy_io.ComfyNode):
                     control_after_generate=True,
                     tooltip="Seed for video generation.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     default="",
-                    force_input=True,
+                    multiline=True,
                     tooltip="An optional text description of undesired elements on an image.",
                     optional=True,
                 ),
-                comfy_io.Custom(PixverseIO.TEMPLATE).Input(
+                IO.Custom(PixverseIO.TEMPLATE).Input(
                     "pixverse_template",
                     tooltip="An optional template to influence style of generation, created by the PixVerse Template node.",
                     optional=True,
                 ),
             ],
-            outputs=[comfy_io.Video.Output()],
+            outputs=[IO.Video.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -314,7 +314,7 @@ class PixverseImageToVideoNode(comfy_io.ComfyNode):
         seed,
         negative_prompt: str = None,
         pixverse_template: int = None,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=False)
         auth = {
             "auth_token": cls.hidden.auth_token_comfy_org,
@@ -377,44 +377,44 @@ class PixverseImageToVideoNode(comfy_io.ComfyNode):
 
         async with aiohttp.ClientSession() as session:
             async with session.get(response_poll.Resp.url) as vid_response:
-                return comfy_io.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read())))
+                return IO.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read())))
 
 
-class PixverseTransitionVideoNode(comfy_io.ComfyNode):
+class PixverseTransitionVideoNode(IO.ComfyNode):
     """
     Generates videos based on prompt and output_size.
     """
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="PixverseTransitionVideoNode",
             display_name="PixVerse Transition Video",
             category="api node/video/PixVerse",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("first_frame"),
-                comfy_io.Image.Input("last_frame"),
-                comfy_io.String.Input(
+                IO.Image.Input("first_frame"),
+                IO.Image.Input("last_frame"),
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt for the video generation",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "quality",
-                    options=[resolution.value for resolution in PixverseQuality],
+                    options=PixverseQuality,
                     default=PixverseQuality.res_540p,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "duration_seconds",
-                    options=[dur.value for dur in PixverseDuration],
+                    options=PixverseDuration,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "motion_mode",
-                    options=[mode.value for mode in PixverseMotionMode],
+                    options=PixverseMotionMode,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
@@ -422,19 +422,19 @@ class PixverseTransitionVideoNode(comfy_io.ComfyNode):
                     control_after_generate=True,
                     tooltip="Seed for video generation.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     default="",
-                    force_input=True,
+                    multiline=True,
                     tooltip="An optional text description of undesired elements on an image.",
                     optional=True,
                 ),
             ],
-            outputs=[comfy_io.Video.Output()],
+            outputs=[IO.Video.Output()],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -450,7 +450,7 @@ class PixverseTransitionVideoNode(comfy_io.ComfyNode):
         motion_mode: str,
         seed,
         negative_prompt: str = None,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=False)
         auth = {
             "auth_token": cls.hidden.auth_token_comfy_org,
@@ -514,12 +514,12 @@ class PixverseTransitionVideoNode(comfy_io.ComfyNode):
 
         async with aiohttp.ClientSession() as session:
             async with session.get(response_poll.Resp.url) as vid_response:
-                return comfy_io.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read())))
+                return IO.NodeOutput(VideoFromFile(BytesIO(await vid_response.content.read())))
 
 
 class PixVerseExtension(ComfyExtension):
     @override
-    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
         return [
             PixverseTextToVideoNode,
             PixverseImageToVideoNode,

comfy_api_nodes/nodes_recraft.py

@@ -35,6 +35,7 @@ from server import PromptServer
 import torch
 from io import BytesIO
 from PIL import UnidentifiedImageError
+import aiohttp
 
 
 async def handle_recraft_file_request(
@@ -82,10 +83,16 @@ async def handle_recraft_file_request(
     return all_bytesio
 
 
-def recraft_multipart_parser(data, parent_key=None, formatter: callable=None, converted_to_check: list[list]=None, is_list=False) -> dict:
+def recraft_multipart_parser(
+    data,
+    parent_key=None,
+    formatter: callable = None,
+    converted_to_check: list[list] = None,
+    is_list: bool = False,
+    return_mode: str = "formdata"  # "dict" | "formdata"
+) -> dict | aiohttp.FormData:
     """
-    Formats data such that multipart/form-data will work with requests library
-    when both files and data are present.
+    Formats data such that multipart/form-data will work with aiohttp library when both files and data are present.
 
     The OpenAI client that Recraft uses has a bizarre way of serializing lists:
 
@@ -103,23 +110,23 @@ def recraft_multipart_parser(data, parent_key=None, formatter: callable=None, co
     # Modification of a function that handled a different type of multipart parsing, big ups:
     # https://gist.github.com/kazqvaizer/4cebebe5db654a414132809f9f88067b
 
-    def handle_converted_lists(data, parent_key, lists_to_check=tuple[list]):
+    def handle_converted_lists(item, parent_key, lists_to_check=tuple[list]):
         # if list already exists exists, just extend list with data
         for check_list in lists_to_check:
             for conv_tuple in check_list:
-                if conv_tuple[0] == parent_key and type(conv_tuple[1]) is list:
-                    conv_tuple[1].append(formatter(data))
+                if conv_tuple[0] == parent_key and isinstance(conv_tuple[1], list):
+                    conv_tuple[1].append(formatter(item))
                     return True
         return False
 
     if converted_to_check is None:
         converted_to_check = []
 
-
+    effective_mode = return_mode if parent_key is None else "dict"
     if formatter is None:
         formatter = lambda v: v  # Multipart representation of value
 
-    if type(data) is not dict:
+    if not isinstance(data, dict):
         # if list already exists exists, just extend list with data
         added = handle_converted_lists(data, parent_key, converted_to_check)
         if added:
@@ -136,15 +143,24 @@ def recraft_multipart_parser(data, parent_key=None, formatter: callable=None, co
 
     for key, value in data.items():
         current_key = key if parent_key is None else f"{parent_key}[{key}]"
-        if type(value) is dict:
+        if isinstance(value, dict):
             converted.extend(recraft_multipart_parser(value, current_key, formatter, next_check).items())
-        elif type(value) is list:
+        elif isinstance(value, list):
             for ind, list_value in enumerate(value):
                 iter_key = f"{current_key}[]"
                 converted.extend(recraft_multipart_parser(list_value, iter_key, formatter, next_check, is_list=True).items())
         else:
             converted.append((current_key, formatter(value)))
 
+    if effective_mode == "formdata":
+        fd = aiohttp.FormData()
+        for k, v in dict(converted).items():
+            if isinstance(v, list):
+                for item in v:
+                    fd.add_field(k, str(item))
+            else:
+                fd.add_field(k, str(v))
+        return fd
     return dict(converted)
 
 

comfy_api_nodes/nodes_rodin.py

@@ -32,20 +32,20 @@ from comfy_api_nodes.apis.client import (
     SynchronousOperation,
     PollingOperation,
 )
-from comfy_api.latest import ComfyExtension, io as comfy_io
+from comfy_api.latest import ComfyExtension, IO
 
 
 COMMON_PARAMETERS = [
-    comfy_io.Int.Input(
+    IO.Int.Input(
         "Seed",
         default=0,
         min=0,
         max=65535,
-        display_mode=comfy_io.NumberDisplay.number,
+        display_mode=IO.NumberDisplay.number,
         optional=True,
     ),
-    comfy_io.Combo.Input("Material_Type", options=["PBR", "Shaded"], default="PBR", optional=True),
-    comfy_io.Combo.Input(
+    IO.Combo.Input("Material_Type", options=["PBR", "Shaded"], default="PBR", optional=True),
+    IO.Combo.Input(
         "Polygon_count",
         options=["4K-Quad", "8K-Quad", "18K-Quad", "50K-Quad", "200K-Triangle"],
         default="18K-Quad",
@@ -172,16 +172,16 @@ async def create_generate_task(
     logging.info("[ Rodin3D API - Submit Jobs ] Submit Generate Task Success!")
     subscription_key = response.jobs.subscription_key
     task_uuid = response.uuid
-    logging.info(f"[ Rodin3D API - Submit Jobs ] UUID: {task_uuid}")
+    logging.info("[ Rodin3D API - Submit Jobs ] UUID: %s", task_uuid)
     return task_uuid, subscription_key
 
 
 def check_rodin_status(response: Rodin3DCheckStatusResponse) -> str:
     all_done = all(job.status == JobStatus.Done for job in response.jobs)
     status_list = [str(job.status) for job in response.jobs]
-    logging.info(f"[ Rodin3D API - CheckStatus ] Generate Status: {status_list}")
+    logging.info("[ Rodin3D API - CheckStatus ] Generate Status: %s", status_list)
     if any(job.status == JobStatus.Failed for job in response.jobs):
-        logging.error(f"[ Rodin3D API - CheckStatus ] Generate Failed: {status_list}, Please try again.")
+        logging.error("[ Rodin3D API - CheckStatus ] Generate Failed: %s, Please try again.", status_list)
         raise Exception("[ Rodin3D API ] Generate Failed, Please Try again.")
     if all_done:
         return "DONE"
@@ -235,7 +235,7 @@ async def download_files(url_list, task_uuid):
             file_path = os.path.join(save_path, file_name)
             if file_path.endswith(".glb"):
                 model_file_path = file_path
-            logging.info(f"[ Rodin3D API - download_files ] Downloading file: {file_path}")
+            logging.info("[ Rodin3D API - download_files ] Downloading file: %s", file_path)
             max_retries = 5
             for attempt in range(max_retries):
                 try:
@@ -246,7 +246,7 @@ async def download_files(url_list, task_uuid):
                                 f.write(chunk)
                     break
                 except Exception as e:
-                    logging.info(f"[ Rodin3D API - download_files ] Error downloading {file_path}:{e}")
+                    logging.info("[ Rodin3D API - download_files ] Error downloading %s:%s", file_path, str(e))
                     if attempt < max_retries - 1:
                         logging.info("Retrying...")
                         await asyncio.sleep(2)
@@ -259,24 +259,24 @@ async def download_files(url_list, task_uuid):
     return model_file_path
 
 
-class Rodin3D_Regular(comfy_io.ComfyNode):
+class Rodin3D_Regular(IO.ComfyNode):
     """Generate 3D Assets using Rodin API"""
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="Rodin3D_Regular",
             display_name="Rodin 3D Generate - Regular Generate",
             category="api node/3d/Rodin",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("Images"),
+                IO.Image.Input("Images"),
                 *COMMON_PARAMETERS,
             ],
-            outputs=[comfy_io.String.Output(display_name="3D Model Path")],
+            outputs=[IO.String.Output(display_name="3D Model Path")],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
             ],
             is_api_node=True,
         )
@@ -288,7 +288,7 @@ class Rodin3D_Regular(comfy_io.ComfyNode):
         Seed,
         Material_Type,
         Polygon_count,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         tier = "Regular"
         num_images = Images.shape[0]
         m_images = []
@@ -312,27 +312,27 @@ class Rodin3D_Regular(comfy_io.ComfyNode):
         download_list = await get_rodin_download_list(task_uuid, auth_kwargs=auth)
         model = await download_files(download_list, task_uuid)
 
-        return comfy_io.NodeOutput(model)
+        return IO.NodeOutput(model)
 
 
-class Rodin3D_Detail(comfy_io.ComfyNode):
+class Rodin3D_Detail(IO.ComfyNode):
     """Generate 3D Assets using Rodin API"""
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="Rodin3D_Detail",
             display_name="Rodin 3D Generate - Detail Generate",
             category="api node/3d/Rodin",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("Images"),
+                IO.Image.Input("Images"),
                 *COMMON_PARAMETERS,
             ],
-            outputs=[comfy_io.String.Output(display_name="3D Model Path")],
+            outputs=[IO.String.Output(display_name="3D Model Path")],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
             ],
             is_api_node=True,
         )
@@ -344,7 +344,7 @@ class Rodin3D_Detail(comfy_io.ComfyNode):
         Seed,
         Material_Type,
         Polygon_count,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         tier = "Detail"
         num_images = Images.shape[0]
         m_images = []
@@ -368,27 +368,27 @@ class Rodin3D_Detail(comfy_io.ComfyNode):
         download_list = await get_rodin_download_list(task_uuid, auth_kwargs=auth)
         model = await download_files(download_list, task_uuid)
 
-        return comfy_io.NodeOutput(model)
+        return IO.NodeOutput(model)
 
 
-class Rodin3D_Smooth(comfy_io.ComfyNode):
+class Rodin3D_Smooth(IO.ComfyNode):
     """Generate 3D Assets using Rodin API"""
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="Rodin3D_Smooth",
             display_name="Rodin 3D Generate - Smooth Generate",
             category="api node/3d/Rodin",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("Images"),
+                IO.Image.Input("Images"),
                 *COMMON_PARAMETERS,
             ],
-            outputs=[comfy_io.String.Output(display_name="3D Model Path")],
+            outputs=[IO.String.Output(display_name="3D Model Path")],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
             ],
             is_api_node=True,
         )
@@ -400,7 +400,7 @@ class Rodin3D_Smooth(comfy_io.ComfyNode):
         Seed,
         Material_Type,
         Polygon_count,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         tier = "Smooth"
         num_images = Images.shape[0]
         m_images = []
@@ -424,34 +424,34 @@ class Rodin3D_Smooth(comfy_io.ComfyNode):
         download_list = await get_rodin_download_list(task_uuid, auth_kwargs=auth)
         model = await download_files(download_list, task_uuid)
 
-        return comfy_io.NodeOutput(model)
+        return IO.NodeOutput(model)
 
 
-class Rodin3D_Sketch(comfy_io.ComfyNode):
+class Rodin3D_Sketch(IO.ComfyNode):
     """Generate 3D Assets using Rodin API"""
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="Rodin3D_Sketch",
             display_name="Rodin 3D Generate - Sketch Generate",
             category="api node/3d/Rodin",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("Images"),
-                comfy_io.Int.Input(
+                IO.Image.Input("Images"),
+                IO.Int.Input(
                     "Seed",
                     default=0,
                     min=0,
                     max=65535,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     optional=True,
                 ),
             ],
-            outputs=[comfy_io.String.Output(display_name="3D Model Path")],
+            outputs=[IO.String.Output(display_name="3D Model Path")],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
             ],
             is_api_node=True,
         )
@@ -461,7 +461,7 @@ class Rodin3D_Sketch(comfy_io.ComfyNode):
         cls,
         Images,
         Seed,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         tier = "Sketch"
         num_images = Images.shape[0]
         m_images = []
@@ -487,42 +487,42 @@ class Rodin3D_Sketch(comfy_io.ComfyNode):
         download_list = await get_rodin_download_list(task_uuid, auth_kwargs=auth)
         model = await download_files(download_list, task_uuid)
 
-        return comfy_io.NodeOutput(model)
+        return IO.NodeOutput(model)
 
 
-class Rodin3D_Gen2(comfy_io.ComfyNode):
+class Rodin3D_Gen2(IO.ComfyNode):
     """Generate 3D Assets using Rodin API"""
 
     @classmethod
-    def define_schema(cls) -> comfy_io.Schema:
-        return comfy_io.Schema(
+    def define_schema(cls) -> IO.Schema:
+        return IO.Schema(
             node_id="Rodin3D_Gen2",
             display_name="Rodin 3D Generate - Gen-2 Generate",
             category="api node/3d/Rodin",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("Images"),
-                comfy_io.Int.Input(
+                IO.Image.Input("Images"),
+                IO.Int.Input(
                     "Seed",
                     default=0,
                     min=0,
                     max=65535,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     optional=True,
                 ),
-                comfy_io.Combo.Input("Material_Type", options=["PBR", "Shaded"], default="PBR", optional=True),
-                comfy_io.Combo.Input(
+                IO.Combo.Input("Material_Type", options=["PBR", "Shaded"], default="PBR", optional=True),
+                IO.Combo.Input(
                     "Polygon_count",
                     options=["4K-Quad", "8K-Quad", "18K-Quad", "50K-Quad", "2K-Triangle", "20K-Triangle", "150K-Triangle", "500K-Triangle"],
                     default="500K-Triangle",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input("TAPose", default=False),
+                IO.Boolean.Input("TAPose", default=False),
             ],
-            outputs=[comfy_io.String.Output(display_name="3D Model Path")],
+            outputs=[IO.String.Output(display_name="3D Model Path")],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
             ],
             is_api_node=True,
         )
@@ -535,7 +535,7 @@ class Rodin3D_Gen2(comfy_io.ComfyNode):
         Material_Type,
         Polygon_count,
         TAPose,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         tier = "Gen-2"
         num_images = Images.shape[0]
         m_images = []
@@ -560,12 +560,12 @@ class Rodin3D_Gen2(comfy_io.ComfyNode):
         download_list = await get_rodin_download_list(task_uuid, auth_kwargs=auth)
         model = await download_files(download_list, task_uuid)
 
-        return comfy_io.NodeOutput(model)
+        return IO.NodeOutput(model)
 
 
 class Rodin3DExtension(ComfyExtension):
     @override
-    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
         return [
             Rodin3D_Regular,
             Rodin3D_Detail,

comfy_api_nodes/nodes_runway.py

@@ -48,7 +48,7 @@ from comfy_api_nodes.apinode_utils import (
     download_url_to_image_tensor,
 )
 from comfy_api.input_impl import VideoFromFile
-from comfy_api.latest import ComfyExtension, io as comfy_io
+from comfy_api.latest import ComfyExtension, IO
 from comfy_api_nodes.util.validation_utils import validate_image_dimensions, validate_image_aspect_ratio
 
 PATH_IMAGE_TO_VIDEO = "/proxy/runway/image_to_video"
@@ -175,11 +175,11 @@ async def generate_video(
     return await download_url_to_video_output(video_url)
 
 
-class RunwayImageToVideoNodeGen3a(comfy_io.ComfyNode):
+class RunwayImageToVideoNodeGen3a(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="RunwayImageToVideoNodeGen3a",
             display_name="Runway Image to Video (Gen3a Turbo)",
             category="api node/video/Runway",
@@ -188,42 +188,42 @@ class RunwayImageToVideoNodeGen3a(comfy_io.ComfyNode):
                         "your input selections will set your generation up for success: "
                         "https://help.runwayml.com/hc/en-us/articles/33927968552339-Creating-with-Act-One-on-Gen-3-Alpha-and-Turbo.",
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Text prompt for the generation",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "start_frame",
                     tooltip="Start frame to be used for the video",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "duration",
-                    options=[model.value for model in Duration],
+                    options=Duration,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "ratio",
-                    options=[model.value for model in RunwayGen3aAspectRatio],
+                    options=RunwayGen3aAspectRatio,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=4294967295,
                     step=1,
                     control_after_generate=True,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Random seed for generation",
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -236,7 +236,7 @@ class RunwayImageToVideoNodeGen3a(comfy_io.ComfyNode):
         duration: str,
         ratio: str,
         seed: int,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, min_length=1)
         validate_image_dimensions(start_frame, max_width=7999, max_height=7999)
         validate_image_aspect_ratio(start_frame, min_aspect_ratio=0.5, max_aspect_ratio=2.0)
@@ -253,7 +253,7 @@ class RunwayImageToVideoNodeGen3a(comfy_io.ComfyNode):
             auth_kwargs=auth_kwargs,
         )
 
-        return comfy_io.NodeOutput(
+        return IO.NodeOutput(
             await generate_video(
                 RunwayImageToVideoRequest(
                     promptText=prompt,
@@ -275,11 +275,11 @@ class RunwayImageToVideoNodeGen3a(comfy_io.ComfyNode):
         )
 
 
-class RunwayImageToVideoNodeGen4(comfy_io.ComfyNode):
+class RunwayImageToVideoNodeGen4(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="RunwayImageToVideoNodeGen4",
             display_name="Runway Image to Video (Gen4 Turbo)",
             category="api node/video/Runway",
@@ -288,42 +288,42 @@ class RunwayImageToVideoNodeGen4(comfy_io.ComfyNode):
                         "your input selections will set your generation up for success: "
                         "https://help.runwayml.com/hc/en-us/articles/37327109429011-Creating-with-Gen-4-Video.",
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Text prompt for the generation",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "start_frame",
                     tooltip="Start frame to be used for the video",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "duration",
-                    options=[model.value for model in Duration],
+                    options=Duration,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "ratio",
-                    options=[model.value for model in RunwayGen4TurboAspectRatio],
+                    options=RunwayGen4TurboAspectRatio,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=4294967295,
                     step=1,
                     control_after_generate=True,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Random seed for generation",
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -336,7 +336,7 @@ class RunwayImageToVideoNodeGen4(comfy_io.ComfyNode):
         duration: str,
         ratio: str,
         seed: int,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, min_length=1)
         validate_image_dimensions(start_frame, max_width=7999, max_height=7999)
         validate_image_aspect_ratio(start_frame, min_aspect_ratio=0.5, max_aspect_ratio=2.0)
@@ -353,7 +353,7 @@ class RunwayImageToVideoNodeGen4(comfy_io.ComfyNode):
             auth_kwargs=auth_kwargs,
         )
 
-        return comfy_io.NodeOutput(
+        return IO.NodeOutput(
             await generate_video(
                 RunwayImageToVideoRequest(
                     promptText=prompt,
@@ -376,11 +376,11 @@ class RunwayImageToVideoNodeGen4(comfy_io.ComfyNode):
         )
 
 
-class RunwayFirstLastFrameNode(comfy_io.ComfyNode):
+class RunwayFirstLastFrameNode(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="RunwayFirstLastFrameNode",
             display_name="Runway First-Last-Frame to Video",
             category="api node/video/Runway",
@@ -392,46 +392,46 @@ class RunwayFirstLastFrameNode(comfy_io.ComfyNode):
                         "will set your generation up for success: "
                         "https://help.runwayml.com/hc/en-us/articles/34170748696595-Creating-with-Keyframes-on-Gen-3.",
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Text prompt for the generation",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "start_frame",
                     tooltip="Start frame to be used for the video",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "end_frame",
                     tooltip="End frame to be used for the video. Supported for gen3a_turbo only.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "duration",
-                    options=[model.value for model in Duration],
+                    options=Duration,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "ratio",
-                    options=[model.value for model in RunwayGen3aAspectRatio],
+                    options=RunwayGen3aAspectRatio,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=4294967295,
                     step=1,
                     control_after_generate=True,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Random seed for generation",
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -445,7 +445,7 @@ class RunwayFirstLastFrameNode(comfy_io.ComfyNode):
         duration: str,
         ratio: str,
         seed: int,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, min_length=1)
         validate_image_dimensions(start_frame, max_width=7999, max_height=7999)
         validate_image_dimensions(end_frame, max_width=7999, max_height=7999)
@@ -467,7 +467,7 @@ class RunwayFirstLastFrameNode(comfy_io.ComfyNode):
         if len(download_urls) != 2:
             raise RunwayApiError("Failed to upload one or more images to comfy api.")
 
-        return comfy_io.NodeOutput(
+        return IO.NodeOutput(
             await generate_video(
                 RunwayImageToVideoRequest(
                     promptText=prompt,
@@ -493,40 +493,40 @@ class RunwayFirstLastFrameNode(comfy_io.ComfyNode):
         )
 
 
-class RunwayTextToImageNode(comfy_io.ComfyNode):
+class RunwayTextToImageNode(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="RunwayTextToImageNode",
             display_name="Runway Text to Image",
             category="api node/image/Runway",
             description="Generate an image from a text prompt using Runway's Gen 4 model. "
                         "You can also include reference image to guide the generation.",
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Text prompt for the generation",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "ratio",
                     options=[model.value for model in RunwayTextToImageAspectRatioEnum],
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "reference_image",
                     tooltip="Optional reference image to guide the generation",
                     optional=True,
                 ),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -537,7 +537,7 @@ class RunwayTextToImageNode(comfy_io.ComfyNode):
         prompt: str,
         ratio: str,
         reference_image: Optional[torch.Tensor] = None,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, min_length=1)
 
         auth_kwargs = {
@@ -588,12 +588,12 @@ class RunwayTextToImageNode(comfy_io.ComfyNode):
         if not final_response.output:
             raise RunwayApiError("Runway task succeeded but no image data found in response.")
 
-        return comfy_io.NodeOutput(await download_url_to_image_tensor(get_image_url_from_task_status(final_response)))
+        return IO.NodeOutput(await download_url_to_image_tensor(get_image_url_from_task_status(final_response)))
 
 
 class RunwayExtension(ComfyExtension):
     @override
-    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
         return [
             RunwayFirstLastFrameNode,
             RunwayImageToVideoNodeGen3a,

comfy_api_nodes/nodes_sora.py

@@ -0,0 +1,175 @@
+from typing import Optional
+from typing_extensions import override
+
+import torch
+from pydantic import BaseModel, Field
+from comfy_api.latest import ComfyExtension, IO
+from comfy_api_nodes.apis.client import (
+    ApiEndpoint,
+    HttpMethod,
+    SynchronousOperation,
+    PollingOperation,
+    EmptyRequest,
+)
+from comfy_api_nodes.util.validation_utils import get_number_of_images
+
+from comfy_api_nodes.apinode_utils import (
+    download_url_to_video_output,
+    tensor_to_bytesio,
+)
+
+class Sora2GenerationRequest(BaseModel):
+    prompt: str = Field(...)
+    model: str = Field(...)
+    seconds: str = Field(...)
+    size: str = Field(...)
+
+
+class Sora2GenerationResponse(BaseModel):
+    id: str = Field(...)
+    error: Optional[dict] = Field(None)
+    status: Optional[str] = Field(None)
+
+
+class OpenAIVideoSora2(IO.ComfyNode):
+    @classmethod
+    def define_schema(cls):
+        return IO.Schema(
+            node_id="OpenAIVideoSora2",
+            display_name="OpenAI Sora - Video",
+            category="api node/video/Sora",
+            description="OpenAI video and audio generation.",
+            inputs=[
+                IO.Combo.Input(
+                    "model",
+                    options=["sora-2", "sora-2-pro"],
+                    default="sora-2",
+                ),
+                IO.String.Input(
+                    "prompt",
+                    multiline=True,
+                    default="",
+                    tooltip="Guiding text; may be empty if an input image is present.",
+                ),
+                IO.Combo.Input(
+                    "size",
+                    options=[
+                        "720x1280",
+                        "1280x720",
+                        "1024x1792",
+                        "1792x1024",
+                    ],
+                    default="1280x720",
+                ),
+                IO.Combo.Input(
+                    "duration",
+                    options=[4, 8, 12],
+                    default=8,
+                ),
+                IO.Image.Input(
+                    "image",
+                    optional=True,
+                ),
+                IO.Int.Input(
+                    "seed",
+                    default=0,
+                    min=0,
+                    max=2147483647,
+                    step=1,
+                    display_mode=IO.NumberDisplay.number,
+                    control_after_generate=True,
+                    optional=True,
+                    tooltip="Seed to determine if node should re-run; "
+                            "actual results are nondeterministic regardless of seed.",
+                ),
+            ],
+            outputs=[
+                IO.Video.Output(),
+            ],
+            hidden=[
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
+            ],
+            is_api_node=True,
+        )
+
+    @classmethod
+    async def execute(
+        cls,
+        model: str,
+        prompt: str,
+        size: str = "1280x720",
+        duration: int = 8,
+        seed: int = 0,
+        image: Optional[torch.Tensor] = None,
+    ):
+        if model == "sora-2" and size not in ("720x1280", "1280x720"):
+            raise ValueError("Invalid size for sora-2 model, only 720x1280 and 1280x720 are supported.")
+        files_input = None
+        if image is not None:
+            if get_number_of_images(image) != 1:
+                raise ValueError("Currently only one input image is supported.")
+            files_input = {"input_reference": ("image.png", tensor_to_bytesio(image), "image/png")}
+        auth = {
+            "auth_token": cls.hidden.auth_token_comfy_org,
+            "comfy_api_key": cls.hidden.api_key_comfy_org,
+        }
+        payload = Sora2GenerationRequest(
+            model=model,
+            prompt=prompt,
+            seconds=str(duration),
+            size=size,
+        )
+        initial_operation = SynchronousOperation(
+            endpoint=ApiEndpoint(
+                path="/proxy/openai/v1/videos",
+                method=HttpMethod.POST,
+                request_model=Sora2GenerationRequest,
+                response_model=Sora2GenerationResponse
+            ),
+            request=payload,
+            files=files_input,
+            auth_kwargs=auth,
+            content_type="multipart/form-data",
+        )
+        initial_response = await initial_operation.execute()
+        if initial_response.error:
+            raise Exception(initial_response.error.message)
+
+        model_time_multiplier = 1 if model == "sora-2" else 2
+        poll_operation = PollingOperation(
+            poll_endpoint=ApiEndpoint(
+                path=f"/proxy/openai/v1/videos/{initial_response.id}",
+                method=HttpMethod.GET,
+                request_model=EmptyRequest,
+                response_model=Sora2GenerationResponse
+            ),
+            completed_statuses=["completed"],
+            failed_statuses=["failed"],
+            status_extractor=lambda x: x.status,
+            auth_kwargs=auth,
+            poll_interval=8.0,
+            max_poll_attempts=160,
+            node_id=cls.hidden.unique_id,
+            estimated_duration=45 * (duration / 4) * model_time_multiplier,
+        )
+        await poll_operation.execute()
+        return IO.NodeOutput(
+            await download_url_to_video_output(
+                f"/proxy/openai/v1/videos/{initial_response.id}/content",
+                auth_kwargs=auth,
+            )
+        )
+
+
+class OpenAISoraExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
+        return [
+            OpenAIVideoSora2,
+        ]
+
+
+async def comfy_entrypoint() -> OpenAISoraExtension:
+    return OpenAISoraExtension()

comfy_api_nodes/nodes_stability.py

@@ -2,7 +2,7 @@ from inspect import cleandoc
 from typing import Optional
 from typing_extensions import override
 
-from comfy_api.latest import ComfyExtension, Input, io as comfy_io
+from comfy_api.latest import ComfyExtension, Input, IO
 from comfy_api_nodes.apis.stability_api import (
     StabilityUpscaleConservativeRequest,
     StabilityUpscaleCreativeRequest,
@@ -56,20 +56,20 @@ def get_async_dummy_status(x: StabilityResultsGetResponse):
     return StabilityPollStatus.in_progress
 
 
-class StabilityStableImageUltraNode(comfy_io.ComfyNode):
+class StabilityStableImageUltraNode(IO.ComfyNode):
     """
     Generates images synchronously based on prompt and resolution.
     """
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="StabilityStableImageUltraNode",
             display_name="Stability AI Stable Image Ultra",
             category="api node/image/Stability AI",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
@@ -80,39 +80,39 @@ class StabilityStableImageUltraNode(comfy_io.ComfyNode):
                                     "is a value between 0 and 1. For example: `The sky was a crisp (blue:0.3) and (green:0.8)`" +
                                     "would convey a sky that was blue and green, but more green than blue.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
-                    options=[x.value for x in StabilityAspectRatio],
-                    default=StabilityAspectRatio.ratio_1_1.value,
+                    options=StabilityAspectRatio,
+                    default=StabilityAspectRatio.ratio_1_1,
                     tooltip="Aspect ratio of generated image.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "style_preset",
                     options=get_stability_style_presets(),
                     tooltip="Optional desired style of generated image.",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=4294967294,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="The random seed used for creating the noise.",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                     optional=True,
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     default="",
                     tooltip="A blurb of text describing what you do not wish to see in the output image. This is an advanced feature.",
                     force_input=True,
                     optional=True,
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "image_denoise",
                     default=0.5,
                     min=0.0,
@@ -123,12 +123,12 @@ class StabilityStableImageUltraNode(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -143,7 +143,7 @@ class StabilityStableImageUltraNode(comfy_io.ComfyNode):
         image: Optional[torch.Tensor] = None,
         negative_prompt: str = "",
         image_denoise: Optional[float] = 0.5,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=False)
         # prepare image binary if image present
         image_binary = None
@@ -193,44 +193,44 @@ class StabilityStableImageUltraNode(comfy_io.ComfyNode):
         image_data = base64.b64decode(response_api.image)
         returned_image = bytesio_to_image_tensor(BytesIO(image_data))
 
-        return comfy_io.NodeOutput(returned_image)
+        return IO.NodeOutput(returned_image)
 
 
-class StabilityStableImageSD_3_5Node(comfy_io.ComfyNode):
+class StabilityStableImageSD_3_5Node(IO.ComfyNode):
     """
     Generates images synchronously based on prompt and resolution.
     """
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="StabilityStableImageSD_3_5Node",
             display_name="Stability AI Stable Diffusion 3.5 Image",
             category="api node/image/Stability AI",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="What you wish to see in the output image. A strong, descriptive prompt that clearly defines elements, colors, and subjects will lead to better results.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
-                    options=[x.value for x in Stability_SD3_5_Model],
+                    options=Stability_SD3_5_Model,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
-                    options=[x.value for x in StabilityAspectRatio],
-                    default=StabilityAspectRatio.ratio_1_1.value,
+                    options=StabilityAspectRatio,
+                    default=StabilityAspectRatio.ratio_1_1,
                     tooltip="Aspect ratio of generated image.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "style_preset",
                     options=get_stability_style_presets(),
                     tooltip="Optional desired style of generated image.",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "cfg_scale",
                     default=4.0,
                     min=1.0,
@@ -238,28 +238,28 @@ class StabilityStableImageSD_3_5Node(comfy_io.ComfyNode):
                     step=0.1,
                     tooltip="How strictly the diffusion process adheres to the prompt text (higher values keep your image closer to your prompt)",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=4294967294,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="The random seed used for creating the noise.",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                     optional=True,
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     default="",
                     tooltip="Keywords of what you do not wish to see in the output image. This is an advanced feature.",
                     force_input=True,
                     optional=True,
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "image_denoise",
                     default=0.5,
                     min=0.0,
@@ -270,12 +270,12 @@ class StabilityStableImageSD_3_5Node(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -292,7 +292,7 @@ class StabilityStableImageSD_3_5Node(comfy_io.ComfyNode):
         image: Optional[torch.Tensor] = None,
         negative_prompt: str = "",
         image_denoise: Optional[float] = 0.5,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=False)
         # prepare image binary if image present
         image_binary = None
@@ -348,30 +348,30 @@ class StabilityStableImageSD_3_5Node(comfy_io.ComfyNode):
         image_data = base64.b64decode(response_api.image)
         returned_image = bytesio_to_image_tensor(BytesIO(image_data))
 
-        return comfy_io.NodeOutput(returned_image)
+        return IO.NodeOutput(returned_image)
 
 
-class StabilityUpscaleConservativeNode(comfy_io.ComfyNode):
+class StabilityUpscaleConservativeNode(IO.ComfyNode):
     """
     Upscale image with minimal alterations to 4K resolution.
     """
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="StabilityUpscaleConservativeNode",
             display_name="Stability AI Upscale Conservative",
             category="api node/image/Stability AI",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("image"),
-                comfy_io.String.Input(
+                IO.Image.Input("image"),
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="What you wish to see in the output image. A strong, descriptive prompt that clearly defines elements, colors, and subjects will lead to better results.",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "creativity",
                     default=0.35,
                     min=0.2,
@@ -379,17 +379,17 @@ class StabilityUpscaleConservativeNode(comfy_io.ComfyNode):
                     step=0.01,
                     tooltip="Controls the likelihood of creating additional details not heavily conditioned by the init image.",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=4294967294,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="The random seed used for creating the noise.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     default="",
                     tooltip="Keywords of what you do not wish to see in the output image. This is an advanced feature.",
@@ -398,12 +398,12 @@ class StabilityUpscaleConservativeNode(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -416,7 +416,7 @@ class StabilityUpscaleConservativeNode(comfy_io.ComfyNode):
         creativity: float,
         seed: int,
         negative_prompt: str = "",
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=False)
         image_binary = tensor_to_bytesio(image, total_pixels=1024*1024).read()
 
@@ -457,30 +457,30 @@ class StabilityUpscaleConservativeNode(comfy_io.ComfyNode):
         image_data = base64.b64decode(response_api.image)
         returned_image = bytesio_to_image_tensor(BytesIO(image_data))
 
-        return comfy_io.NodeOutput(returned_image)
+        return IO.NodeOutput(returned_image)
 
 
-class StabilityUpscaleCreativeNode(comfy_io.ComfyNode):
+class StabilityUpscaleCreativeNode(IO.ComfyNode):
     """
     Upscale image with minimal alterations to 4K resolution.
     """
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="StabilityUpscaleCreativeNode",
             display_name="Stability AI Upscale Creative",
             category="api node/image/Stability AI",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("image"),
-                comfy_io.String.Input(
+                IO.Image.Input("image"),
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="What you wish to see in the output image. A strong, descriptive prompt that clearly defines elements, colors, and subjects will lead to better results.",
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "creativity",
                     default=0.3,
                     min=0.1,
@@ -488,22 +488,22 @@ class StabilityUpscaleCreativeNode(comfy_io.ComfyNode):
                     step=0.01,
                     tooltip="Controls the likelihood of creating additional details not heavily conditioned by the init image.",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "style_preset",
                     options=get_stability_style_presets(),
                     tooltip="Optional desired style of generated image.",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=4294967294,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="The random seed used for creating the noise.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     default="",
                     tooltip="Keywords of what you do not wish to see in the output image. This is an advanced feature.",
@@ -512,12 +512,12 @@ class StabilityUpscaleCreativeNode(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -531,7 +531,7 @@ class StabilityUpscaleCreativeNode(comfy_io.ComfyNode):
         style_preset: str,
         seed: int,
         negative_prompt: str = "",
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, strip_whitespace=False)
         image_binary = tensor_to_bytesio(image, total_pixels=1024*1024).read()
 
@@ -591,37 +591,37 @@ class StabilityUpscaleCreativeNode(comfy_io.ComfyNode):
         image_data = base64.b64decode(response_poll.result)
         returned_image = bytesio_to_image_tensor(BytesIO(image_data))
 
-        return comfy_io.NodeOutput(returned_image)
+        return IO.NodeOutput(returned_image)
 
 
-class StabilityUpscaleFastNode(comfy_io.ComfyNode):
+class StabilityUpscaleFastNode(IO.ComfyNode):
     """
     Quickly upscales an image via Stability API call to 4x its original size; intended for upscaling low-quality/compressed images.
     """
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="StabilityUpscaleFastNode",
             display_name="Stability AI Upscale Fast",
             category="api node/image/Stability AI",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Image.Input("image"),
+                IO.Image.Input("image"),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
 
     @classmethod
-    async def execute(cls, image: torch.Tensor) -> comfy_io.NodeOutput:
+    async def execute(cls, image: torch.Tensor) -> IO.NodeOutput:
         image_binary = tensor_to_bytesio(image, total_pixels=4096*4096).read()
 
         files = {
@@ -653,26 +653,26 @@ class StabilityUpscaleFastNode(comfy_io.ComfyNode):
         image_data = base64.b64decode(response_api.image)
         returned_image = bytesio_to_image_tensor(BytesIO(image_data))
 
-        return comfy_io.NodeOutput(returned_image)
+        return IO.NodeOutput(returned_image)
 
 
-class StabilityTextToAudio(comfy_io.ComfyNode):
+class StabilityTextToAudio(IO.ComfyNode):
     """Generates high-quality music and sound effects from text descriptions."""
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="StabilityTextToAudio",
             display_name="Stability AI Text To Audio",
             category="api node/audio/Stability AI",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=["stable-audio-2.5"],
                 ),
-                comfy_io.String.Input("prompt", multiline=True, default=""),
-                comfy_io.Int.Input(
+                IO.String.Input("prompt", multiline=True, default=""),
+                IO.Int.Input(
                     "duration",
                     default=190,
                     min=1,
@@ -681,18 +681,18 @@ class StabilityTextToAudio(comfy_io.ComfyNode):
                     tooltip="Controls the duration in seconds of the generated audio.",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=4294967294,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="The random seed used for generation.",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "steps",
                     default=8,
                     min=4,
@@ -703,18 +703,18 @@ class StabilityTextToAudio(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Audio.Output(),
+                IO.Audio.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
 
     @classmethod
-    async def execute(cls, model: str, prompt: str, duration: int, seed: int, steps: int) -> comfy_io.NodeOutput:
+    async def execute(cls, model: str, prompt: str, duration: int, seed: int, steps: int) -> IO.NodeOutput:
         validate_string(prompt, max_length=10000)
         payload = StabilityTextToAudioRequest(prompt=prompt, model=model, duration=duration, seed=seed, steps=steps)
         operation = SynchronousOperation(
@@ -734,27 +734,27 @@ class StabilityTextToAudio(comfy_io.ComfyNode):
         response_api = await operation.execute()
         if not response_api.audio:
             raise ValueError("No audio file was received in response.")
-        return comfy_io.NodeOutput(audio_bytes_to_audio_input(base64.b64decode(response_api.audio)))
+        return IO.NodeOutput(audio_bytes_to_audio_input(base64.b64decode(response_api.audio)))
 
 
-class StabilityAudioToAudio(comfy_io.ComfyNode):
+class StabilityAudioToAudio(IO.ComfyNode):
     """Transforms existing audio samples into new high-quality compositions using text instructions."""
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="StabilityAudioToAudio",
             display_name="Stability AI Audio To Audio",
             category="api node/audio/Stability AI",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=["stable-audio-2.5"],
                 ),
-                comfy_io.String.Input("prompt", multiline=True, default=""),
-                comfy_io.Audio.Input("audio", tooltip="Audio must be between 6 and 190 seconds long."),
-                comfy_io.Int.Input(
+                IO.String.Input("prompt", multiline=True, default=""),
+                IO.Audio.Input("audio", tooltip="Audio must be between 6 and 190 seconds long."),
+                IO.Int.Input(
                     "duration",
                     default=190,
                     min=1,
@@ -763,18 +763,18 @@ class StabilityAudioToAudio(comfy_io.ComfyNode):
                     tooltip="Controls the duration in seconds of the generated audio.",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=4294967294,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="The random seed used for generation.",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "steps",
                     default=8,
                     min=4,
@@ -783,24 +783,24 @@ class StabilityAudioToAudio(comfy_io.ComfyNode):
                     tooltip="Controls the number of sampling steps.",
                     optional=True,
                 ),
-                comfy_io.Float.Input(
+                IO.Float.Input(
                     "strength",
                     default=1,
                     min=0.01,
                     max=1.0,
                     step=0.01,
-                    display_mode=comfy_io.NumberDisplay.slider,
+                    display_mode=IO.NumberDisplay.slider,
                     tooltip="Parameter controls how much influence the audio parameter has on the generated audio.",
                     optional=True,
                 ),
             ],
             outputs=[
-                comfy_io.Audio.Output(),
+                IO.Audio.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -808,7 +808,7 @@ class StabilityAudioToAudio(comfy_io.ComfyNode):
     @classmethod
     async def execute(
         cls, model: str, prompt: str, audio: Input.Audio, duration: int, seed: int, steps: int, strength: float
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, max_length=10000)
         validate_audio_duration(audio, 6, 190)
         payload = StabilityAudioToAudioRequest(
@@ -832,27 +832,27 @@ class StabilityAudioToAudio(comfy_io.ComfyNode):
         response_api = await operation.execute()
         if not response_api.audio:
             raise ValueError("No audio file was received in response.")
-        return comfy_io.NodeOutput(audio_bytes_to_audio_input(base64.b64decode(response_api.audio)))
+        return IO.NodeOutput(audio_bytes_to_audio_input(base64.b64decode(response_api.audio)))
 
 
-class StabilityAudioInpaint(comfy_io.ComfyNode):
+class StabilityAudioInpaint(IO.ComfyNode):
     """Transforms part of existing audio sample using text instructions."""
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="StabilityAudioInpaint",
             display_name="Stability AI Audio Inpaint",
             category="api node/audio/Stability AI",
             description=cleandoc(cls.__doc__ or ""),
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=["stable-audio-2.5"],
                 ),
-                comfy_io.String.Input("prompt", multiline=True, default=""),
-                comfy_io.Audio.Input("audio", tooltip="Audio must be between 6 and 190 seconds long."),
-                comfy_io.Int.Input(
+                IO.String.Input("prompt", multiline=True, default=""),
+                IO.Audio.Input("audio", tooltip="Audio must be between 6 and 190 seconds long."),
+                IO.Int.Input(
                     "duration",
                     default=190,
                     min=1,
@@ -861,18 +861,18 @@ class StabilityAudioInpaint(comfy_io.ComfyNode):
                     tooltip="Controls the duration in seconds of the generated audio.",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=4294967294,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="The random seed used for generation.",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "steps",
                     default=8,
                     min=4,
@@ -881,7 +881,7 @@ class StabilityAudioInpaint(comfy_io.ComfyNode):
                     tooltip="Controls the number of sampling steps.",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "mask_start",
                     default=30,
                     min=0,
@@ -889,7 +889,7 @@ class StabilityAudioInpaint(comfy_io.ComfyNode):
                     step=1,
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "mask_end",
                     default=190,
                     min=0,
@@ -899,12 +899,12 @@ class StabilityAudioInpaint(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Audio.Output(),
+                IO.Audio.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -920,7 +920,7 @@ class StabilityAudioInpaint(comfy_io.ComfyNode):
         steps: int,
         mask_start: int,
         mask_end: int,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_string(prompt, max_length=10000)
         if mask_end <= mask_start:
             raise ValueError(f"Value of mask_end({mask_end}) should be greater then mask_start({mask_start})")
@@ -953,12 +953,12 @@ class StabilityAudioInpaint(comfy_io.ComfyNode):
         response_api = await operation.execute()
         if not response_api.audio:
             raise ValueError("No audio file was received in response.")
-        return comfy_io.NodeOutput(audio_bytes_to_audio_input(base64.b64decode(response_api.audio)))
+        return IO.NodeOutput(audio_bytes_to_audio_input(base64.b64decode(response_api.audio)))
 
 
 class StabilityExtension(ComfyExtension):
     @override
-    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
         return [
             StabilityStableImageUltraNode,
             StabilityStableImageSD_3_5Node,

comfy_api_nodes/nodes_veo2.py

@@ -6,7 +6,7 @@ from io import BytesIO
 from typing import Optional
 from typing_extensions import override
 
-from comfy_api.latest import ComfyExtension, io as comfy_io
+from comfy_api.latest import ComfyExtension, IO
 from comfy_api.input_impl.video_types import VideoFromFile
 from comfy_api_nodes.apis import (
     VeoGenVidRequest,
@@ -27,6 +27,13 @@ from comfy_api_nodes.apinode_utils import (
 )
 
 AVERAGE_DURATION_VIDEO_GEN = 32
+MODELS_MAP = {
+    "veo-2.0-generate-001": "veo-2.0-generate-001",
+    "veo-3.1-generate": "veo-3.1-generate-preview",
+    "veo-3.1-fast-generate": "veo-3.1-fast-generate-preview",
+    "veo-3.0-generate-001": "veo-3.0-generate-001",
+    "veo-3.0-fast-generate-001": "veo-3.0-fast-generate-001",
+}
 
 def convert_image_to_base64(image: torch.Tensor):
     if image is None:
@@ -51,7 +58,7 @@ def get_video_url_from_response(poll_response: VeoGenVidPollResponse) -> Optiona
     return None
 
 
-class VeoVideoGenerationNode(comfy_io.ComfyNode):
+class VeoVideoGenerationNode(IO.ComfyNode):
     """
     Generates videos from text prompts using Google's Veo API.
 
@@ -61,71 +68,71 @@ class VeoVideoGenerationNode(comfy_io.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="VeoVideoGenerationNode",
             display_name="Google Veo 2 Video Generation",
             category="api node/video/Veo",
             description="Generates videos from text prompts using Google's Veo 2 API",
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Text description of the video",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
                     options=["16:9", "9:16"],
                     default="16:9",
                     tooltip="Aspect ratio of the output video",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     multiline=True,
                     default="",
                     tooltip="Negative text prompt to guide what to avoid in the video",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "duration_seconds",
                     default=5,
                     min=5,
                     max=8,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Duration of the output video in seconds",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "enhance_prompt",
                     default=True,
                     tooltip="Whether to enhance the prompt with AI assistance",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "person_generation",
                     options=["ALLOW", "BLOCK"],
                     default="ALLOW",
                     tooltip="Whether to allow generating people in the video",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=0xFFFFFFFF,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed for video generation (0 for random)",
                     optional=True,
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                     tooltip="Optional reference image to guide video generation",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=["veo-2.0-generate-001"],
                     default="veo-2.0-generate-001",
@@ -134,12 +141,12 @@ class VeoVideoGenerationNode(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -158,6 +165,7 @@ class VeoVideoGenerationNode(comfy_io.ComfyNode):
         model="veo-2.0-generate-001",
         generate_audio=False,
     ):
+        model = MODELS_MAP[model]
         # Prepare the instances for the request
         instances = []
 
@@ -215,7 +223,7 @@ class VeoVideoGenerationNode(comfy_io.ComfyNode):
         initial_response = await initial_operation.execute()
         operation_name = initial_response.name
 
-        logging.info(f"Veo generation started with operation name: {operation_name}")
+        logging.info("Veo generation started with operation name: %s", operation_name)
 
         # Define status extractor function
         def status_extractor(response):
@@ -302,7 +310,7 @@ class VeoVideoGenerationNode(comfy_io.ComfyNode):
         video_io = BytesIO(video_data)
 
         # Return VideoFromFile object
-        return comfy_io.NodeOutput(VideoFromFile(video_io))
+        return IO.NodeOutput(VideoFromFile(video_io))
 
 
 class Veo3VideoGenerationNode(VeoVideoGenerationNode):
@@ -319,78 +327,80 @@ class Veo3VideoGenerationNode(VeoVideoGenerationNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="Veo3VideoGenerationNode",
             display_name="Google Veo 3 Video Generation",
             category="api node/video/Veo",
             description="Generates videos from text prompts using Google's Veo 3 API",
             inputs=[
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Text description of the video",
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
                     options=["16:9", "9:16"],
                     default="16:9",
                     tooltip="Aspect ratio of the output video",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     multiline=True,
                     default="",
                     tooltip="Negative text prompt to guide what to avoid in the video",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "duration_seconds",
                     default=8,
                     min=8,
                     max=8,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Duration of the output video in seconds (Veo 3 only supports 8 seconds)",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "enhance_prompt",
                     default=True,
                     tooltip="Whether to enhance the prompt with AI assistance",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "person_generation",
                     options=["ALLOW", "BLOCK"],
                     default="ALLOW",
                     tooltip="Whether to allow generating people in the video",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=0xFFFFFFFF,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed for video generation (0 for random)",
                     optional=True,
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                     tooltip="Optional reference image to guide video generation",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
-                    options=["veo-3.0-generate-001", "veo-3.0-fast-generate-001"],
+                    options=[
+                        "veo-3.1-generate", "veo-3.1-fast-generate", "veo-3.0-generate-001", "veo-3.0-fast-generate-001"
+                    ],
                     default="veo-3.0-generate-001",
                     tooltip="Veo 3 model to use for video generation",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "generate_audio",
                     default=False,
                     tooltip="Generate audio for the video. Supported by all Veo 3 models.",
@@ -398,12 +408,12 @@ class Veo3VideoGenerationNode(VeoVideoGenerationNode):
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -411,7 +421,7 @@ class Veo3VideoGenerationNode(VeoVideoGenerationNode):
 
 class VeoExtension(ComfyExtension):
     @override
-    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
         return [
             VeoVideoGenerationNode,
             Veo3VideoGenerationNode,

comfy_api_nodes/nodes_vidu.py

@@ -6,7 +6,7 @@ from typing_extensions import override
 import torch
 from pydantic import BaseModel, Field
 
-from comfy_api.latest import ComfyExtension, io as comfy_io
+from comfy_api.latest import ComfyExtension, IO
 from comfy_api_nodes.util.validation_utils import (
     validate_aspect_ratio_closeness,
     validate_image_dimensions,
@@ -161,77 +161,77 @@ async def execute_task(
     )
 
 
-class ViduTextToVideoNode(comfy_io.ComfyNode):
+class ViduTextToVideoNode(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="ViduTextToVideoNode",
             display_name="Vidu Text To Video Generation",
             category="api node/video/Vidu",
             description="Generate video from text prompt",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
-                    options=[model.value for model in VideoModelName],
-                    default=VideoModelName.vidu_q1.value,
+                    options=VideoModelName,
+                    default=VideoModelName.vidu_q1,
                     tooltip="Model name",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     tooltip="A textual description for video generation",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "duration",
                     default=5,
                     min=5,
                     max=5,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Duration of the output video in seconds",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed for video generation (0 for random)",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
-                    options=[model.value for model in AspectRatio],
-                    default=AspectRatio.r_16_9.value,
+                    options=AspectRatio,
+                    default=AspectRatio.r_16_9,
                     tooltip="The aspect ratio of the output video",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
-                    options=[model.value for model in Resolution],
-                    default=Resolution.r_1080p.value,
+                    options=Resolution,
+                    default=Resolution.r_1080p,
                     tooltip="Supported values may vary by model & duration",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "movement_amplitude",
-                    options=[model.value for model in MovementAmplitude],
-                    default=MovementAmplitude.auto.value,
+                    options=MovementAmplitude,
+                    default=MovementAmplitude.auto,
                     tooltip="The movement amplitude of objects in the frame",
                     optional=True,
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -246,7 +246,7 @@ class ViduTextToVideoNode(comfy_io.ComfyNode):
         aspect_ratio: str,
         resolution: str,
         movement_amplitude: str,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         if not prompt:
             raise ValueError("The prompt field is required and cannot be empty.")
         payload = TaskCreationRequest(
@@ -263,79 +263,79 @@ class ViduTextToVideoNode(comfy_io.ComfyNode):
             "comfy_api_key": cls.hidden.api_key_comfy_org,
         }
         results = await execute_task(VIDU_TEXT_TO_VIDEO, auth, payload, 320, cls.hidden.unique_id)
-        return comfy_io.NodeOutput(await download_url_to_video_output(get_video_from_response(results).url))
+        return IO.NodeOutput(await download_url_to_video_output(get_video_from_response(results).url))
 
 
-class ViduImageToVideoNode(comfy_io.ComfyNode):
+class ViduImageToVideoNode(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="ViduImageToVideoNode",
             display_name="Vidu Image To Video Generation",
             category="api node/video/Vidu",
             description="Generate video from image and optional prompt",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
-                    options=[model.value for model in VideoModelName],
-                    default=VideoModelName.vidu_q1.value,
+                    options=VideoModelName,
+                    default=VideoModelName.vidu_q1,
                     tooltip="Model name",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                     tooltip="An image to be used as the start frame of the generated video",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="A textual description for video generation",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "duration",
                     default=5,
                     min=5,
                     max=5,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Duration of the output video in seconds",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed for video generation (0 for random)",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
-                    options=[model.value for model in Resolution],
-                    default=Resolution.r_1080p.value,
+                    options=Resolution,
+                    default=Resolution.r_1080p,
                     tooltip="Supported values may vary by model & duration",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "movement_amplitude",
-                    options=[model.value for model in MovementAmplitude],
+                    options=MovementAmplitude,
                     default=MovementAmplitude.auto.value,
                     tooltip="The movement amplitude of objects in the frame",
                     optional=True,
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -350,7 +350,7 @@ class ViduImageToVideoNode(comfy_io.ComfyNode):
         seed: int,
         resolution: str,
         movement_amplitude: str,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         if get_number_of_images(image) > 1:
             raise ValueError("Only one input image is allowed.")
         validate_image_aspect_ratio_range(image, (1, 4), (4, 1))
@@ -373,70 +373,70 @@ class ViduImageToVideoNode(comfy_io.ComfyNode):
             auth_kwargs=auth,
         )
         results = await execute_task(VIDU_IMAGE_TO_VIDEO, auth, payload, 120, cls.hidden.unique_id)
-        return comfy_io.NodeOutput(await download_url_to_video_output(get_video_from_response(results).url))
+        return IO.NodeOutput(await download_url_to_video_output(get_video_from_response(results).url))
 
 
-class ViduReferenceVideoNode(comfy_io.ComfyNode):
+class ViduReferenceVideoNode(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="ViduReferenceVideoNode",
             display_name="Vidu Reference To Video Generation",
             category="api node/video/Vidu",
             description="Generate video from multiple images and prompt",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
-                    options=[model.value for model in VideoModelName],
-                    default=VideoModelName.vidu_q1.value,
+                    options=VideoModelName,
+                    default=VideoModelName.vidu_q1,
                     tooltip="Model name",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "images",
                     tooltip="Images to use as references to generate a video with consistent subjects (max 7 images).",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     tooltip="A textual description for video generation",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "duration",
                     default=5,
                     min=5,
                     max=5,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Duration of the output video in seconds",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed for video generation (0 for random)",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "aspect_ratio",
-                    options=[model.value for model in AspectRatio],
-                    default=AspectRatio.r_16_9.value,
+                    options=AspectRatio,
+                    default=AspectRatio.r_16_9,
                     tooltip="The aspect ratio of the output video",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
                     options=[model.value for model in Resolution],
                     default=Resolution.r_1080p.value,
                     tooltip="Supported values may vary by model & duration",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "movement_amplitude",
                     options=[model.value for model in MovementAmplitude],
                     default=MovementAmplitude.auto.value,
@@ -445,12 +445,12 @@ class ViduReferenceVideoNode(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -466,7 +466,7 @@ class ViduReferenceVideoNode(comfy_io.ComfyNode):
         aspect_ratio: str,
         resolution: str,
         movement_amplitude: str,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         if not prompt:
             raise ValueError("The prompt field is required and cannot be empty.")
         a = get_number_of_images(images)
@@ -495,68 +495,68 @@ class ViduReferenceVideoNode(comfy_io.ComfyNode):
             auth_kwargs=auth,
         )
         results = await execute_task(VIDU_REFERENCE_VIDEO, auth, payload, 120, cls.hidden.unique_id)
-        return comfy_io.NodeOutput(await download_url_to_video_output(get_video_from_response(results).url))
+        return IO.NodeOutput(await download_url_to_video_output(get_video_from_response(results).url))
 
 
-class ViduStartEndToVideoNode(comfy_io.ComfyNode):
+class ViduStartEndToVideoNode(IO.ComfyNode):
 
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="ViduStartEndToVideoNode",
             display_name="Vidu Start End To Video Generation",
             category="api node/video/Vidu",
             description="Generate a video from start and end frames and a prompt",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=[model.value for model in VideoModelName],
                     default=VideoModelName.vidu_q1.value,
                     tooltip="Model name",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "first_frame",
                     tooltip="Start frame",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "end_frame",
                     tooltip="End frame",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     tooltip="A textual description for video generation",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "duration",
                     default=5,
                     min=5,
                     max=5,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Duration of the output video in seconds",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed for video generation (0 for random)",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
                     options=[model.value for model in Resolution],
                     default=Resolution.r_1080p.value,
                     tooltip="Supported values may vary by model & duration",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "movement_amplitude",
                     options=[model.value for model in MovementAmplitude],
                     default=MovementAmplitude.auto.value,
@@ -565,12 +565,12 @@ class ViduStartEndToVideoNode(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -586,7 +586,7 @@ class ViduStartEndToVideoNode(comfy_io.ComfyNode):
         seed: int,
         resolution: str,
         movement_amplitude: str,
-    ) -> comfy_io.NodeOutput:
+    ) -> IO.NodeOutput:
         validate_aspect_ratio_closeness(first_frame, end_frame, min_rel=0.8, max_rel=1.25, strict=False)
         payload = TaskCreationRequest(
             model_name=model,
@@ -605,12 +605,12 @@ class ViduStartEndToVideoNode(comfy_io.ComfyNode):
             for frame in (first_frame, end_frame)
         ]
         results = await execute_task(VIDU_START_END_VIDEO, auth, payload, 96, cls.hidden.unique_id)
-        return comfy_io.NodeOutput(await download_url_to_video_output(get_video_from_response(results).url))
+        return IO.NodeOutput(await download_url_to_video_output(get_video_from_response(results).url))
 
 
 class ViduExtension(ComfyExtension):
     @override
-    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
         return [
             ViduTextToVideoNode,
             ViduImageToVideoNode,

comfy_api_nodes/nodes_wan.py

@@ -4,7 +4,7 @@ from typing_extensions import override
 
 import torch
 from pydantic import BaseModel, Field
-from comfy_api.latest import ComfyExtension, Input, io as comfy_io
+from comfy_api.latest import ComfyExtension, Input, IO
 from comfy_api_nodes.apis.client import (
     ApiEndpoint,
     HttpMethod,
@@ -195,35 +195,35 @@ async def process_task(
     ).execute()
 
 
-class WanTextToImageApi(comfy_io.ComfyNode):
+class WanTextToImageApi(IO.ComfyNode):
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="WanTextToImageApi",
             display_name="Wan Text to Image",
             category="api node/image/Wan",
             description="Generates image based on text prompt.",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=["wan2.5-t2i-preview"],
                     default="wan2.5-t2i-preview",
                     tooltip="Model to use.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt used to describe the elements and visual features, supports English/Chinese.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     multiline=True,
                     default="",
                     tooltip="Negative text prompt to guide what to avoid.",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "width",
                     default=1024,
                     min=768,
@@ -231,7 +231,7 @@ class WanTextToImageApi(comfy_io.ComfyNode):
                     step=32,
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "height",
                     default=1024,
                     min=768,
@@ -239,24 +239,24 @@ class WanTextToImageApi(comfy_io.ComfyNode):
                     step=32,
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed to use for generation.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "prompt_extend",
                     default=True,
                     tooltip="Whether to enhance the prompt with AI assistance.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "watermark",
                     default=True,
                     tooltip="Whether to add an \"AI generated\" watermark to the result.",
@@ -264,12 +264,12 @@ class WanTextToImageApi(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -309,36 +309,36 @@ class WanTextToImageApi(comfy_io.ComfyNode):
             estimated_duration=9,
             poll_interval=3,
         )
-        return comfy_io.NodeOutput(await download_url_to_image_tensor(str(response.output.results[0].url)))
+        return IO.NodeOutput(await download_url_to_image_tensor(str(response.output.results[0].url)))
 
 
-class WanImageToImageApi(comfy_io.ComfyNode):
+class WanImageToImageApi(IO.ComfyNode):
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="WanImageToImageApi",
             display_name="Wan Image to Image",
             category="api node/image/Wan",
             description="Generates an image from one or two input images and a text prompt. "
                         "The output image is currently fixed at 1.6 MP; its aspect ratio matches the input image(s).",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=["wan2.5-i2i-preview"],
                     default="wan2.5-i2i-preview",
                     tooltip="Model to use.",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                     tooltip="Single-image editing or multi-image fusion, maximum 2 images.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt used to describe the elements and visual features, supports English/Chinese.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     multiline=True,
                     default="",
@@ -346,7 +346,7 @@ class WanImageToImageApi(comfy_io.ComfyNode):
                     optional=True,
                 ),
                 # redo this later as an optional combo of recommended resolutions
-                # comfy_io.Int.Input(
+                # IO.Int.Input(
                 #     "width",
                 #     default=1280,
                 #     min=384,
@@ -354,7 +354,7 @@ class WanImageToImageApi(comfy_io.ComfyNode):
                 #     step=16,
                 #     optional=True,
                 # ),
-                # comfy_io.Int.Input(
+                # IO.Int.Input(
                 #     "height",
                 #     default=1280,
                 #     min=384,
@@ -362,18 +362,18 @@ class WanImageToImageApi(comfy_io.ComfyNode):
                 #     step=16,
                 #     optional=True,
                 # ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed to use for generation.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "watermark",
                     default=True,
                     tooltip="Whether to add an \"AI generated\" watermark to the result.",
@@ -381,12 +381,12 @@ class WanImageToImageApi(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Image.Output(),
+                IO.Image.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -431,38 +431,38 @@ class WanImageToImageApi(comfy_io.ComfyNode):
             estimated_duration=42,
             poll_interval=3,
         )
-        return comfy_io.NodeOutput(await download_url_to_image_tensor(str(response.output.results[0].url)))
+        return IO.NodeOutput(await download_url_to_image_tensor(str(response.output.results[0].url)))
 
 
-class WanTextToVideoApi(comfy_io.ComfyNode):
+class WanTextToVideoApi(IO.ComfyNode):
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="WanTextToVideoApi",
             display_name="Wan Text to Video",
             category="api node/video/Wan",
             description="Generates video based on text prompt.",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=["wan2.5-t2v-preview"],
                     default="wan2.5-t2v-preview",
                     tooltip="Model to use.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt used to describe the elements and visual features, supports English/Chinese.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     multiline=True,
                     default="",
                     tooltip="Negative text prompt to guide what to avoid.",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "size",
                     options=[
                         "480p: 1:1 (624x624)",
@@ -482,45 +482,45 @@ class WanTextToVideoApi(comfy_io.ComfyNode):
                     default="480p: 1:1 (624x624)",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "duration",
                     default=5,
                     min=5,
                     max=10,
                     step=5,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Available durations: 5 and 10 seconds",
                     optional=True,
                 ),
-                comfy_io.Audio.Input(
+                IO.Audio.Input(
                     "audio",
                     optional=True,
                     tooltip="Audio must contain a clear, loud voice, without extraneous noise, background music.",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed to use for generation.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "generate_audio",
                     default=False,
                     optional=True,
                     tooltip="If there is no audio input, generate audio automatically.",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "prompt_extend",
                     default=True,
                     tooltip="Whether to enhance the prompt with AI assistance.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "watermark",
                     default=True,
                     tooltip="Whether to add an \"AI generated\" watermark to the result.",
@@ -528,12 +528,12 @@ class WanTextToVideoApi(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -582,41 +582,41 @@ class WanTextToVideoApi(comfy_io.ComfyNode):
             estimated_duration=120 * int(duration / 5),
             poll_interval=6,
         )
-        return comfy_io.NodeOutput(await download_url_to_video_output(response.output.video_url))
+        return IO.NodeOutput(await download_url_to_video_output(response.output.video_url))
 
 
-class WanImageToVideoApi(comfy_io.ComfyNode):
+class WanImageToVideoApi(IO.ComfyNode):
     @classmethod
     def define_schema(cls):
-        return comfy_io.Schema(
+        return IO.Schema(
             node_id="WanImageToVideoApi",
             display_name="Wan Image to Video",
             category="api node/video/Wan",
             description="Generates video based on the first frame and text prompt.",
             inputs=[
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "model",
                     options=["wan2.5-i2v-preview"],
                     default="wan2.5-i2v-preview",
                     tooltip="Model to use.",
                 ),
-                comfy_io.Image.Input(
+                IO.Image.Input(
                     "image",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "prompt",
                     multiline=True,
                     default="",
                     tooltip="Prompt used to describe the elements and visual features, supports English/Chinese.",
                 ),
-                comfy_io.String.Input(
+                IO.String.Input(
                     "negative_prompt",
                     multiline=True,
                     default="",
                     tooltip="Negative text prompt to guide what to avoid.",
                     optional=True,
                 ),
-                comfy_io.Combo.Input(
+                IO.Combo.Input(
                     "resolution",
                     options=[
                         "480P",
@@ -626,45 +626,45 @@ class WanImageToVideoApi(comfy_io.ComfyNode):
                     default="480P",
                     optional=True,
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "duration",
                     default=5,
                     min=5,
                     max=10,
                     step=5,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     tooltip="Available durations: 5 and 10 seconds",
                     optional=True,
                 ),
-                comfy_io.Audio.Input(
+                IO.Audio.Input(
                     "audio",
                     optional=True,
                     tooltip="Audio must contain a clear, loud voice, without extraneous noise, background music.",
                 ),
-                comfy_io.Int.Input(
+                IO.Int.Input(
                     "seed",
                     default=0,
                     min=0,
                     max=2147483647,
                     step=1,
-                    display_mode=comfy_io.NumberDisplay.number,
+                    display_mode=IO.NumberDisplay.number,
                     control_after_generate=True,
                     tooltip="Seed to use for generation.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "generate_audio",
                     default=False,
                     optional=True,
                     tooltip="If there is no audio input, generate audio automatically.",
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "prompt_extend",
                     default=True,
                     tooltip="Whether to enhance the prompt with AI assistance.",
                     optional=True,
                 ),
-                comfy_io.Boolean.Input(
+                IO.Boolean.Input(
                     "watermark",
                     default=True,
                     tooltip="Whether to add an \"AI generated\" watermark to the result.",
@@ -672,12 +672,12 @@ class WanImageToVideoApi(comfy_io.ComfyNode):
                 ),
             ],
             outputs=[
-                comfy_io.Video.Output(),
+                IO.Video.Output(),
             ],
             hidden=[
-                comfy_io.Hidden.auth_token_comfy_org,
-                comfy_io.Hidden.api_key_comfy_org,
-                comfy_io.Hidden.unique_id,
+                IO.Hidden.auth_token_comfy_org,
+                IO.Hidden.api_key_comfy_org,
+                IO.Hidden.unique_id,
             ],
             is_api_node=True,
         )
@@ -731,12 +731,12 @@ class WanImageToVideoApi(comfy_io.ComfyNode):
             estimated_duration=120 * int(duration / 5),
             poll_interval=6,
         )
-        return comfy_io.NodeOutput(await download_url_to_video_output(response.output.video_url))
+        return IO.NodeOutput(await download_url_to_video_output(response.output.video_url))
 
 
 class WanApiExtension(ComfyExtension):
     @override
-    async def get_node_list(self) -> list[type[comfy_io.ComfyNode]]:
+    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
         return [
             WanTextToImageApi,
             WanImageToImageApi,

comfy_extras/nodes_audio.py

@@ -142,9 +142,10 @@ def save_audio(self, audio, filename_prefix="ComfyUI", format="flac", prompt=Non
         for key, value in metadata.items():
             output_container.metadata[key] = value
 
+        layout = 'mono' if waveform.shape[0] == 1 else 'stereo'
         # Set up the output stream with appropriate properties
         if format == "opus":
-            out_stream = output_container.add_stream("libopus", rate=sample_rate)
+            out_stream = output_container.add_stream("libopus", rate=sample_rate, layout=layout)
             if quality == "64k":
                 out_stream.bit_rate = 64000
             elif quality == "96k":
@@ -156,7 +157,7 @@ def save_audio(self, audio, filename_prefix="ComfyUI", format="flac", prompt=Non
             elif quality == "320k":
                 out_stream.bit_rate = 320000
         elif format == "mp3":
-            out_stream = output_container.add_stream("libmp3lame", rate=sample_rate)
+            out_stream = output_container.add_stream("libmp3lame", rate=sample_rate, layout=layout)
             if quality == "V0":
                 #TODO i would really love to support V3 and V5 but there doesn't seem to be a way to set the qscale level, the property below is a bool
                 out_stream.codec_context.qscale = 1
@@ -165,9 +166,9 @@ def save_audio(self, audio, filename_prefix="ComfyUI", format="flac", prompt=Non
             elif quality == "320k":
                 out_stream.bit_rate = 320000
         else: #format == "flac":
-            out_stream = output_container.add_stream("flac", rate=sample_rate)
+            out_stream = output_container.add_stream("flac", rate=sample_rate, layout=layout)
 
-        frame = av.AudioFrame.from_ndarray(waveform.movedim(0, 1).reshape(1, -1).float().numpy(), format='flt', layout='mono' if waveform.shape[0] == 1 else 'stereo')
+        frame = av.AudioFrame.from_ndarray(waveform.movedim(0, 1).reshape(1, -1).float().numpy(), format='flt', layout=layout)
         frame.sample_rate = sample_rate
         frame.pts = 0
         output_container.mux(out_stream.encode(frame))

comfy_extras/nodes_compositing.py

@@ -1,6 +1,9 @@
 import torch
 import comfy.utils
 from enum import Enum
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
+
 
 def resize_mask(mask, shape):
     return torch.nn.functional.interpolate(mask.reshape((-1, 1, mask.shape[-2], mask.shape[-1])), size=(shape[0], shape[1]), mode="bilinear").squeeze(1)
@@ -101,24 +104,28 @@ def porter_duff_composite(src_image: torch.Tensor, src_alpha: torch.Tensor, dst_
     return out_image, out_alpha
 
 
-class PorterDuffImageComposite:
+class PorterDuffImageComposite(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {
-            "required": {
-                "source": ("IMAGE",),
-                "source_alpha": ("MASK",),
-                "destination": ("IMAGE",),
-                "destination_alpha": ("MASK",),
-                "mode": ([mode.name for mode in PorterDuffMode], {"default": PorterDuffMode.DST.name}),
-            },
-        }
+    def define_schema(cls):
+        return io.Schema(
+            node_id="PorterDuffImageComposite",
+            display_name="Porter-Duff Image Composite",
+            category="mask/compositing",
+            inputs=[
+                io.Image.Input("source"),
+                io.Mask.Input("source_alpha"),
+                io.Image.Input("destination"),
+                io.Mask.Input("destination_alpha"),
+                io.Combo.Input("mode", options=[mode.name for mode in PorterDuffMode], default=PorterDuffMode.DST.name),
+            ],
+            outputs=[
+                io.Image.Output(),
+                io.Mask.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("IMAGE", "MASK")
-    FUNCTION = "composite"
-    CATEGORY = "mask/compositing"
-
-    def composite(self, source: torch.Tensor, source_alpha: torch.Tensor, destination: torch.Tensor, destination_alpha: torch.Tensor, mode):
+    @classmethod
+    def execute(cls, source: torch.Tensor, source_alpha: torch.Tensor, destination: torch.Tensor, destination_alpha: torch.Tensor, mode) -> io.NodeOutput:
         batch_size = min(len(source), len(source_alpha), len(destination), len(destination_alpha))
         out_images = []
         out_alphas = []
@@ -150,45 +157,48 @@ class PorterDuffImageComposite:
             out_images.append(out_image)
             out_alphas.append(out_alpha.squeeze(2))
 
-        result = (torch.stack(out_images), torch.stack(out_alphas))
-        return result
+        return io.NodeOutput(torch.stack(out_images), torch.stack(out_alphas))
 
 
-class SplitImageWithAlpha:
+class SplitImageWithAlpha(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {
-                "required": {
-                    "image": ("IMAGE",),
-                }
-        }
+    def define_schema(cls):
+        return io.Schema(
+            node_id="SplitImageWithAlpha",
+            display_name="Split Image with Alpha",
+            category="mask/compositing",
+            inputs=[
+                io.Image.Input("image"),
+            ],
+            outputs=[
+                io.Image.Output(),
+                io.Mask.Output(),
+            ],
+        )
 
-    CATEGORY = "mask/compositing"
-    RETURN_TYPES = ("IMAGE", "MASK")
-    FUNCTION = "split_image_with_alpha"
-
-    def split_image_with_alpha(self, image: torch.Tensor):
+    @classmethod
+    def execute(cls, image: torch.Tensor) -> io.NodeOutput:
         out_images = [i[:,:,:3] for i in image]
         out_alphas = [i[:,:,3] if i.shape[2] > 3 else torch.ones_like(i[:,:,0]) for i in image]
-        result = (torch.stack(out_images), 1.0 - torch.stack(out_alphas))
-        return result
+        return io.NodeOutput(torch.stack(out_images), 1.0 - torch.stack(out_alphas))
 
 
-class JoinImageWithAlpha:
+class JoinImageWithAlpha(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {
-                "required": {
-                    "image": ("IMAGE",),
-                    "alpha": ("MASK",),
-                }
-        }
+    def define_schema(cls):
+        return io.Schema(
+            node_id="JoinImageWithAlpha",
+            display_name="Join Image with Alpha",
+            category="mask/compositing",
+            inputs=[
+                io.Image.Input("image"),
+                io.Mask.Input("alpha"),
+            ],
+            outputs=[io.Image.Output()],
+        )
 
-    CATEGORY = "mask/compositing"
-    RETURN_TYPES = ("IMAGE",)
-    FUNCTION = "join_image_with_alpha"
-
-    def join_image_with_alpha(self, image: torch.Tensor, alpha: torch.Tensor):
+    @classmethod
+    def execute(cls, image: torch.Tensor, alpha: torch.Tensor) -> io.NodeOutput:
         batch_size = min(len(image), len(alpha))
         out_images = []
 
@@ -196,19 +206,18 @@ class JoinImageWithAlpha:
         for i in range(batch_size):
            out_images.append(torch.cat((image[i][:,:,:3], alpha[i].unsqueeze(2)), dim=2))
 
-        result = (torch.stack(out_images),)
-        return result
+        return io.NodeOutput(torch.stack(out_images))
 
 
-NODE_CLASS_MAPPINGS = {
-    "PorterDuffImageComposite": PorterDuffImageComposite,
-    "SplitImageWithAlpha": SplitImageWithAlpha,
-    "JoinImageWithAlpha": JoinImageWithAlpha,
-}
+class CompositingExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            PorterDuffImageComposite,
+            SplitImageWithAlpha,
+            JoinImageWithAlpha,
+        ]
 
 
-NODE_DISPLAY_NAME_MAPPINGS = {
-    "PorterDuffImageComposite": "Porter-Duff Image Composite",
-    "SplitImageWithAlpha": "Split Image with Alpha",
-    "JoinImageWithAlpha": "Join Image with Alpha",
-}
+async def comfy_entrypoint() -> CompositingExtension:
+    return CompositingExtension()

comfy_extras/nodes_controlnet.py

@@ -1,20 +1,26 @@
 from comfy.cldm.control_types import UNION_CONTROLNET_TYPES
 import nodes
 import comfy.utils
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
-class SetUnionControlNetType:
+class SetUnionControlNetType(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"control_net": ("CONTROL_NET", ),
-                             "type": (["auto"] + list(UNION_CONTROLNET_TYPES.keys()),)
-                             }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="SetUnionControlNetType",
+            category="conditioning/controlnet",
+            inputs=[
+                io.ControlNet.Input("control_net"),
+                io.Combo.Input("type", options=["auto"] + list(UNION_CONTROLNET_TYPES.keys())),
+            ],
+            outputs=[
+                io.ControlNet.Output(),
+            ],
+        )
 
-    CATEGORY = "conditioning/controlnet"
-    RETURN_TYPES = ("CONTROL_NET",)
-
-    FUNCTION = "set_controlnet_type"
-
-    def set_controlnet_type(self, control_net, type):
+    @classmethod
+    def execute(cls, control_net, type) -> io.NodeOutput:
         control_net = control_net.copy()
         type_number = UNION_CONTROLNET_TYPES.get(type, -1)
         if type_number >= 0:
@@ -22,27 +28,36 @@ class SetUnionControlNetType:
         else:
             control_net.set_extra_arg("control_type", [])
 
-        return (control_net,)
+        return io.NodeOutput(control_net)
 
-class ControlNetInpaintingAliMamaApply(nodes.ControlNetApplyAdvanced):
+    set_controlnet_type = execute  # TODO: remove
+
+
+class ControlNetInpaintingAliMamaApply(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"positive": ("CONDITIONING", ),
-                             "negative": ("CONDITIONING", ),
-                             "control_net": ("CONTROL_NET", ),
-                             "vae": ("VAE", ),
-                             "image": ("IMAGE", ),
-                             "mask": ("MASK", ),
-                             "strength": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 10.0, "step": 0.01}),
-                             "start_percent": ("FLOAT", {"default": 0.0, "min": 0.0, "max": 1.0, "step": 0.001}),
-                             "end_percent": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.001})
-                             }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="ControlNetInpaintingAliMamaApply",
+            category="conditioning/controlnet",
+            inputs=[
+                io.Conditioning.Input("positive"),
+                io.Conditioning.Input("negative"),
+                io.ControlNet.Input("control_net"),
+                io.Vae.Input("vae"),
+                io.Image.Input("image"),
+                io.Mask.Input("mask"),
+                io.Float.Input("strength", default=1.0, min=0.0, max=10.0, step=0.01),
+                io.Float.Input("start_percent", default=0.0, min=0.0, max=1.0, step=0.001),
+                io.Float.Input("end_percent", default=1.0, min=0.0, max=1.0, step=0.001),
+            ],
+            outputs=[
+                io.Conditioning.Output(display_name="positive"),
+                io.Conditioning.Output(display_name="negative"),
+            ],
+        )
 
-    FUNCTION = "apply_inpaint_controlnet"
-
-    CATEGORY = "conditioning/controlnet"
-
-    def apply_inpaint_controlnet(self, positive, negative, control_net, vae, image, mask, strength, start_percent, end_percent):
+    @classmethod
+    def execute(cls, positive, negative, control_net, vae, image, mask, strength, start_percent, end_percent) -> io.NodeOutput:
         extra_concat = []
         if control_net.concat_mask:
             mask = 1.0 - mask.reshape((-1, 1, mask.shape[-2], mask.shape[-1]))
@@ -50,11 +65,20 @@ class ControlNetInpaintingAliMamaApply(nodes.ControlNetApplyAdvanced):
             image = image * mask_apply.movedim(1, -1).repeat(1, 1, 1, image.shape[3])
             extra_concat = [mask]
 
-        return self.apply_controlnet(positive, negative, control_net, image, strength, start_percent, end_percent, vae=vae, extra_concat=extra_concat)
+        result = nodes.ControlNetApplyAdvanced().apply_controlnet(positive, negative, control_net, image, strength, start_percent, end_percent, vae=vae, extra_concat=extra_concat)
+        return io.NodeOutput(result[0], result[1])
+
+    apply_inpaint_controlnet = execute  # TODO: remove
 
 
+class ControlNetExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            SetUnionControlNetType,
+            ControlNetInpaintingAliMamaApply,
+        ]
 
-NODE_CLASS_MAPPINGS = {
-    "SetUnionControlNetType": SetUnionControlNetType,
-    "ControlNetInpaintingAliMamaApply": ControlNetInpaintingAliMamaApply,
-}
+
+async def comfy_entrypoint() -> ControlNetExtension:
+    return ControlNetExtension()

comfy_extras/nodes_easycache.py

@@ -244,6 +244,8 @@ class EasyCacheHolder:
             self.total_steps_skipped += 1
         batch_offset = x.shape[0] // len(uuids)
         for i, uuid in enumerate(uuids):
+            # slice out only what is relevant to this cond
+            batch_slice = [slice(i*batch_offset,(i+1)*batch_offset)]
             # if cached dims don't match x dims, cut off excess and hope for the best (cosmos world2video)
             if x.shape[1:] != self.uuid_cache_diffs[uuid].shape[1:]:
                 if not self.allow_mismatch:
@@ -261,9 +263,8 @@ class EasyCacheHolder:
                             slicing.append(slice(None, dim_u))
                     else:
                         slicing.append(slice(None))
-                slicing = [slice(i*batch_offset,(i+1)*batch_offset)] + slicing
-                x = x[slicing]
-            x += self.uuid_cache_diffs[uuid].to(x.device)
+                batch_slice = batch_slice + slicing
+            x[batch_slice] += self.uuid_cache_diffs[uuid].to(x.device)
         return x
 
     def update_cache_diff(self, output: torch.Tensor, x: torch.Tensor, uuids: list[UUID]):

comfy_extras/nodes_eps.py

@@ -1,5 +1,7 @@
+import torch
 from typing_extensions import override
 
+from comfy.k_diffusion.sampling import sigma_to_half_log_snr
 from comfy_api.latest import ComfyExtension, io
 
 
@@ -63,12 +65,105 @@ class EpsilonScaling(io.ComfyNode):
         return io.NodeOutput(model_clone)
 
 
+def compute_tsr_rescaling_factor(
+    snr: torch.Tensor, tsr_k: float, tsr_variance: float
+) -> torch.Tensor:
+    """Compute the rescaling score ratio in Temporal Score Rescaling.
+
+    See equation (6) in https://arxiv.org/pdf/2510.01184v1.
+    """
+    posinf_mask = torch.isposinf(snr)
+    rescaling_factor = (snr * tsr_variance + 1) / (snr * tsr_variance / tsr_k + 1)
+    return torch.where(posinf_mask, tsr_k, rescaling_factor) # when snr → inf, r = tsr_k
+
+
+class TemporalScoreRescaling(io.ComfyNode):
+    @classmethod
+    def define_schema(cls):
+        return io.Schema(
+            node_id="TemporalScoreRescaling",
+            display_name="TSR - Temporal Score Rescaling",
+            category="model_patches/unet",
+            inputs=[
+                io.Model.Input("model"),
+                io.Float.Input(
+                    "tsr_k",
+                    tooltip=(
+                        "Controls the rescaling strength.\n"
+                        "Lower k produces more detailed results; higher k produces smoother results in image generation. Setting k = 1 disables rescaling."
+                    ),
+                    default=0.95,
+                    min=0.01,
+                    max=100.0,
+                    step=0.001,
+                    display_mode=io.NumberDisplay.number,
+                ),
+                io.Float.Input(
+                    "tsr_sigma",
+                    tooltip=(
+                        "Controls how early rescaling takes effect.\n"
+                        "Larger values take effect earlier."
+                    ),
+                    default=1.0,
+                    min=0.01,
+                    max=100.0,
+                    step=0.001,
+                    display_mode=io.NumberDisplay.number,
+                ),
+            ],
+            outputs=[
+                io.Model.Output(
+                    display_name="patched_model",
+                ),
+            ],
+            description=(
+                "[Post-CFG Function]\n"
+                "TSR - Temporal Score Rescaling (2510.01184)\n\n"
+                "Rescaling the model's score or noise to steer the sampling diversity.\n"
+            ),
+        )
+
+    @classmethod
+    def execute(cls, model, tsr_k, tsr_sigma) -> io.NodeOutput:
+        tsr_variance = tsr_sigma**2
+
+        def temporal_score_rescaling(args):
+            denoised = args["denoised"]
+            x = args["input"]
+            sigma = args["sigma"]
+            curr_model = args["model"]
+
+            # No rescaling (r = 1) or no noise
+            if tsr_k == 1 or sigma == 0:
+                return denoised
+
+            model_sampling = curr_model.current_patcher.get_model_object("model_sampling")
+            half_log_snr = sigma_to_half_log_snr(sigma, model_sampling)
+            snr = (2 * half_log_snr).exp()
+
+            # No rescaling needed (r = 1)
+            if snr == 0:
+                return denoised
+
+            rescaling_r = compute_tsr_rescaling_factor(snr, tsr_k, tsr_variance)
+
+            # Derived from scaled_denoised = (x - r * sigma * noise) / alpha
+            alpha = sigma * half_log_snr.exp()
+            return torch.lerp(x / alpha, denoised, rescaling_r)
+
+        m = model.clone()
+        m.set_model_sampler_post_cfg_function(temporal_score_rescaling)
+        return io.NodeOutput(m)
+
+
 class EpsilonScalingExtension(ComfyExtension):
     @override
     async def get_node_list(self) -> list[type[io.ComfyNode]]:
         return [
             EpsilonScaling,
+            TemporalScoreRescaling,
         ]
 
+
 async def comfy_entrypoint() -> EpsilonScalingExtension:
     return EpsilonScalingExtension()

comfy_extras/nodes_flux.py

@@ -1,60 +1,80 @@
 import node_helpers
 import comfy.utils
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
-class CLIPTextEncodeFlux:
+
+class CLIPTextEncodeFlux(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-            "clip": ("CLIP", ),
-            "clip_l": ("STRING", {"multiline": True, "dynamicPrompts": True}),
-            "t5xxl": ("STRING", {"multiline": True, "dynamicPrompts": True}),
-            "guidance": ("FLOAT", {"default": 3.5, "min": 0.0, "max": 100.0, "step": 0.1}),
-            }}
-    RETURN_TYPES = ("CONDITIONING",)
-    FUNCTION = "encode"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="CLIPTextEncodeFlux",
+            category="advanced/conditioning/flux",
+            inputs=[
+                io.Clip.Input("clip"),
+                io.String.Input("clip_l", multiline=True, dynamic_prompts=True),
+                io.String.Input("t5xxl", multiline=True, dynamic_prompts=True),
+                io.Float.Input("guidance", default=3.5, min=0.0, max=100.0, step=0.1),
+            ],
+            outputs=[
+                io.Conditioning.Output(),
+            ],
+        )
 
-    CATEGORY = "advanced/conditioning/flux"
-
-    def encode(self, clip, clip_l, t5xxl, guidance):
+    @classmethod
+    def execute(cls, clip, clip_l, t5xxl, guidance) -> io.NodeOutput:
         tokens = clip.tokenize(clip_l)
         tokens["t5xxl"] = clip.tokenize(t5xxl)["t5xxl"]
 
-        return (clip.encode_from_tokens_scheduled(tokens, add_dict={"guidance": guidance}), )
+        return io.NodeOutput(clip.encode_from_tokens_scheduled(tokens, add_dict={"guidance": guidance}))
 
-class FluxGuidance:
+    encode = execute  # TODO: remove
+
+
+class FluxGuidance(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-            "conditioning": ("CONDITIONING", ),
-            "guidance": ("FLOAT", {"default": 3.5, "min": 0.0, "max": 100.0, "step": 0.1}),
-            }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="FluxGuidance",
+            category="advanced/conditioning/flux",
+            inputs=[
+                io.Conditioning.Input("conditioning"),
+                io.Float.Input("guidance", default=3.5, min=0.0, max=100.0, step=0.1),
+            ],
+            outputs=[
+                io.Conditioning.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("CONDITIONING",)
-    FUNCTION = "append"
-
-    CATEGORY = "advanced/conditioning/flux"
-
-    def append(self, conditioning, guidance):
+    @classmethod
+    def execute(cls, conditioning, guidance) -> io.NodeOutput:
         c = node_helpers.conditioning_set_values(conditioning, {"guidance": guidance})
-        return (c, )
+        return io.NodeOutput(c)
+
+    append = execute  # TODO: remove
 
 
-class FluxDisableGuidance:
+class FluxDisableGuidance(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-            "conditioning": ("CONDITIONING", ),
-            }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="FluxDisableGuidance",
+            category="advanced/conditioning/flux",
+            description="This node completely disables the guidance embed on Flux and Flux like models",
+            inputs=[
+                io.Conditioning.Input("conditioning"),
+            ],
+            outputs=[
+                io.Conditioning.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("CONDITIONING",)
-    FUNCTION = "append"
-
-    CATEGORY = "advanced/conditioning/flux"
-    DESCRIPTION = "This node completely disables the guidance embed on Flux and Flux like models"
-
-    def append(self, conditioning):
+    @classmethod
+    def execute(cls, conditioning) -> io.NodeOutput:
         c = node_helpers.conditioning_set_values(conditioning, {"guidance": None})
-        return (c, )
+        return io.NodeOutput(c)
+
+    append = execute  # TODO: remove
 
 
 PREFERED_KONTEXT_RESOLUTIONS = [
@@ -78,52 +98,73 @@ PREFERED_KONTEXT_RESOLUTIONS = [
 ]
 
 
-class FluxKontextImageScale:
+class FluxKontextImageScale(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"image": ("IMAGE", ),
-                            },
-               }
+    def define_schema(cls):
+        return io.Schema(
+            node_id="FluxKontextImageScale",
+            category="advanced/conditioning/flux",
+            description="This node resizes the image to one that is more optimal for flux kontext.",
+            inputs=[
+                io.Image.Input("image"),
+            ],
+            outputs=[
+                io.Image.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("IMAGE",)
-    FUNCTION = "scale"
-
-    CATEGORY = "advanced/conditioning/flux"
-    DESCRIPTION = "This node resizes the image to one that is more optimal for flux kontext."
-
-    def scale(self, image):
+    @classmethod
+    def execute(cls, image) -> io.NodeOutput:
         width = image.shape[2]
         height = image.shape[1]
         aspect_ratio = width / height
         _, width, height = min((abs(aspect_ratio - w / h), w, h) for w, h in PREFERED_KONTEXT_RESOLUTIONS)
         image = comfy.utils.common_upscale(image.movedim(-1, 1), width, height, "lanczos", "center").movedim(1, -1)
-        return (image, )
+        return io.NodeOutput(image)
+
+    scale = execute  # TODO: remove
 
 
-class FluxKontextMultiReferenceLatentMethod:
+class FluxKontextMultiReferenceLatentMethod(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-            "conditioning": ("CONDITIONING", ),
-            "reference_latents_method": (("offset", "index", "uxo/uno"), ),
-            }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="FluxKontextMultiReferenceLatentMethod",
+            category="advanced/conditioning/flux",
+            inputs=[
+                io.Conditioning.Input("conditioning"),
+                io.Combo.Input(
+                    "reference_latents_method",
+                    options=["offset", "index", "uxo/uno"],
+                ),
+            ],
+            outputs=[
+                io.Conditioning.Output(),
+            ],
+            is_experimental=True,
+        )
 
-    RETURN_TYPES = ("CONDITIONING",)
-    FUNCTION = "append"
-    EXPERIMENTAL = True
-
-    CATEGORY = "advanced/conditioning/flux"
-
-    def append(self, conditioning, reference_latents_method):
+    @classmethod
+    def execute(cls, conditioning, reference_latents_method) -> io.NodeOutput:
         if "uxo" in reference_latents_method or "uso" in reference_latents_method:
             reference_latents_method = "uxo"
         c = node_helpers.conditioning_set_values(conditioning, {"reference_latents_method": reference_latents_method})
-        return (c, )
+        return io.NodeOutput(c)
 
-NODE_CLASS_MAPPINGS = {
-    "CLIPTextEncodeFlux": CLIPTextEncodeFlux,
-    "FluxGuidance": FluxGuidance,
-    "FluxDisableGuidance": FluxDisableGuidance,
-    "FluxKontextImageScale": FluxKontextImageScale,
-    "FluxKontextMultiReferenceLatentMethod": FluxKontextMultiReferenceLatentMethod,
-}
+    append = execute  # TODO: remove
+
+
+class FluxExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            CLIPTextEncodeFlux,
+            FluxGuidance,
+            FluxDisableGuidance,
+            FluxKontextImageScale,
+            FluxKontextMultiReferenceLatentMethod,
+        ]
+
+
+async def comfy_entrypoint() -> FluxExtension:
+    return FluxExtension()

comfy_extras/nodes_hunyuan.py

@@ -2,42 +2,60 @@ import nodes
 import node_helpers
 import torch
 import comfy.model_management
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
 
-class CLIPTextEncodeHunyuanDiT:
+class CLIPTextEncodeHunyuanDiT(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-            "clip": ("CLIP", ),
-            "bert": ("STRING", {"multiline": True, "dynamicPrompts": True}),
-            "mt5xl": ("STRING", {"multiline": True, "dynamicPrompts": True}),
-            }}
-    RETURN_TYPES = ("CONDITIONING",)
-    FUNCTION = "encode"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="CLIPTextEncodeHunyuanDiT",
+            category="advanced/conditioning",
+            inputs=[
+                io.Clip.Input("clip"),
+                io.String.Input("bert", multiline=True, dynamic_prompts=True),
+                io.String.Input("mt5xl", multiline=True, dynamic_prompts=True),
+            ],
+            outputs=[
+                io.Conditioning.Output(),
+            ],
+        )
 
-    CATEGORY = "advanced/conditioning"
-
-    def encode(self, clip, bert, mt5xl):
+    @classmethod
+    def execute(cls, clip, bert, mt5xl) -> io.NodeOutput:
         tokens = clip.tokenize(bert)
         tokens["mt5xl"] = clip.tokenize(mt5xl)["mt5xl"]
 
-        return (clip.encode_from_tokens_scheduled(tokens), )
+        return io.NodeOutput(clip.encode_from_tokens_scheduled(tokens))
 
-class EmptyHunyuanLatentVideo:
+    encode = execute  # TODO: remove
+
+
+class EmptyHunyuanLatentVideo(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "width": ("INT", {"default": 848, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 16}),
-                              "height": ("INT", {"default": 480, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 16}),
-                              "length": ("INT", {"default": 25, "min": 1, "max": nodes.MAX_RESOLUTION, "step": 4}),
-                              "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096})}}
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "generate"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="EmptyHunyuanLatentVideo",
+            category="latent/video",
+            inputs=[
+                io.Int.Input("width", default=848, min=16, max=nodes.MAX_RESOLUTION, step=16),
+                io.Int.Input("height", default=480, min=16, max=nodes.MAX_RESOLUTION, step=16),
+                io.Int.Input("length", default=25, min=1, max=nodes.MAX_RESOLUTION, step=4),
+                io.Int.Input("batch_size", default=1, min=1, max=4096),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    CATEGORY = "latent/video"
-
-    def generate(self, width, height, length, batch_size=1):
+    @classmethod
+    def execute(cls, width, height, length, batch_size=1) -> io.NodeOutput:
         latent = torch.zeros([batch_size, 16, ((length - 1) // 4) + 1, height // 8, width // 8], device=comfy.model_management.intermediate_device())
-        return ({"samples":latent}, )
+        return io.NodeOutput({"samples":latent})
+
+    generate = execute  # TODO: remove
+
 
 PROMPT_TEMPLATE_ENCODE_VIDEO_I2V = (
     "<|start_header_id|>system<|end_header_id|>\n\n<image>\nDescribe the video by detailing the following aspects according to the reference image: "
@@ -50,45 +68,61 @@ PROMPT_TEMPLATE_ENCODE_VIDEO_I2V = (
     "<|start_header_id|>assistant<|end_header_id|>\n\n"
 )
 
-class TextEncodeHunyuanVideo_ImageToVideo:
+class TextEncodeHunyuanVideo_ImageToVideo(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-            "clip": ("CLIP", ),
-            "clip_vision_output": ("CLIP_VISION_OUTPUT", ),
-            "prompt": ("STRING", {"multiline": True, "dynamicPrompts": True}),
-            "image_interleave": ("INT", {"default": 2, "min": 1, "max": 512, "tooltip": "How much the image influences things vs the text prompt. Higher number means more influence from the text prompt."}),
-            }}
-    RETURN_TYPES = ("CONDITIONING",)
-    FUNCTION = "encode"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="TextEncodeHunyuanVideo_ImageToVideo",
+            category="advanced/conditioning",
+            inputs=[
+                io.Clip.Input("clip"),
+                io.ClipVisionOutput.Input("clip_vision_output"),
+                io.String.Input("prompt", multiline=True, dynamic_prompts=True),
+                io.Int.Input(
+                    "image_interleave",
+                    default=2,
+                    min=1,
+                    max=512,
+                    tooltip="How much the image influences things vs the text prompt. Higher number means more influence from the text prompt.",
+                ),
+            ],
+            outputs=[
+                io.Conditioning.Output(),
+            ],
+        )
 
-    CATEGORY = "advanced/conditioning"
-
-    def encode(self, clip, clip_vision_output, prompt, image_interleave):
+    @classmethod
+    def execute(cls, clip, clip_vision_output, prompt, image_interleave) -> io.NodeOutput:
         tokens = clip.tokenize(prompt, llama_template=PROMPT_TEMPLATE_ENCODE_VIDEO_I2V, image_embeds=clip_vision_output.mm_projected, image_interleave=image_interleave)
-        return (clip.encode_from_tokens_scheduled(tokens), )
+        return io.NodeOutput(clip.encode_from_tokens_scheduled(tokens))
 
-class HunyuanImageToVideo:
+    encode = execute  # TODO: remove
+
+
+class HunyuanImageToVideo(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"positive": ("CONDITIONING", ),
-                             "vae": ("VAE", ),
-                             "width": ("INT", {"default": 848, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 16}),
-                             "height": ("INT", {"default": 480, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 16}),
-                             "length": ("INT", {"default": 53, "min": 1, "max": nodes.MAX_RESOLUTION, "step": 4}),
-                             "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096}),
-                             "guidance_type": (["v1 (concat)", "v2 (replace)", "custom"], )
-                },
-                "optional": {"start_image": ("IMAGE", ),
-                }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="HunyuanImageToVideo",
+            category="conditioning/video_models",
+            inputs=[
+                io.Conditioning.Input("positive"),
+                io.Vae.Input("vae"),
+                io.Int.Input("width", default=848, min=16, max=nodes.MAX_RESOLUTION, step=16),
+                io.Int.Input("height", default=480, min=16, max=nodes.MAX_RESOLUTION, step=16),
+                io.Int.Input("length", default=53, min=1, max=nodes.MAX_RESOLUTION, step=4),
+                io.Int.Input("batch_size", default=1, min=1, max=4096),
+                io.Combo.Input("guidance_type", options=["v1 (concat)", "v2 (replace)", "custom"]),
+                io.Image.Input("start_image", optional=True),
+            ],
+            outputs=[
+                io.Conditioning.Output(display_name="positive"),
+                io.Latent.Output(display_name="latent"),
+            ],
+        )
 
-    RETURN_TYPES = ("CONDITIONING", "LATENT")
-    RETURN_NAMES = ("positive", "latent")
-    FUNCTION = "encode"
-
-    CATEGORY = "conditioning/video_models"
-
-    def encode(self, positive, vae, width, height, length, batch_size, guidance_type, start_image=None):
+    @classmethod
+    def execute(cls, positive, vae, width, height, length, batch_size, guidance_type, start_image=None) -> io.NodeOutput:
         latent = torch.zeros([batch_size, 16, ((length - 1) // 4) + 1, height // 8, width // 8], device=comfy.model_management.intermediate_device())
         out_latent = {}
 
@@ -111,51 +145,76 @@ class HunyuanImageToVideo:
             positive = node_helpers.conditioning_set_values(positive, cond)
 
         out_latent["samples"] = latent
-        return (positive, out_latent)
+        return io.NodeOutput(positive, out_latent)
 
-class EmptyHunyuanImageLatent:
+    encode = execute  # TODO: remove
+
+
+class EmptyHunyuanImageLatent(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "width": ("INT", {"default": 2048, "min": 64, "max": nodes.MAX_RESOLUTION, "step": 32}),
-                              "height": ("INT", {"default": 2048, "min": 64, "max": nodes.MAX_RESOLUTION, "step": 32}),
-                              "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096})}}
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "generate"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="EmptyHunyuanImageLatent",
+            category="latent",
+            inputs=[
+                io.Int.Input("width", default=2048, min=64, max=nodes.MAX_RESOLUTION, step=32),
+                io.Int.Input("height", default=2048, min=64, max=nodes.MAX_RESOLUTION, step=32),
+                io.Int.Input("batch_size", default=1, min=1, max=4096),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    CATEGORY = "latent"
-
-    def generate(self, width, height, batch_size=1):
+    @classmethod
+    def execute(cls, width, height, batch_size=1) -> io.NodeOutput:
         latent = torch.zeros([batch_size, 64, height // 32, width // 32], device=comfy.model_management.intermediate_device())
-        return ({"samples":latent}, )
+        return io.NodeOutput({"samples":latent})
 
-class HunyuanRefinerLatent:
+    generate = execute  # TODO: remove
+
+
+class HunyuanRefinerLatent(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"positive": ("CONDITIONING", ),
-                             "negative": ("CONDITIONING", ),
-                             "latent": ("LATENT", ),
-                             "noise_augmentation": ("FLOAT", {"default": 0.10, "min": 0.0, "max": 1.0, "step": 0.01}),
-                             }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="HunyuanRefinerLatent",
+            inputs=[
+                io.Conditioning.Input("positive"),
+                io.Conditioning.Input("negative"),
+                io.Latent.Input("latent"),
+                io.Float.Input("noise_augmentation", default=0.10, min=0.0, max=1.0, step=0.01),
 
-    RETURN_TYPES = ("CONDITIONING", "CONDITIONING", "LATENT")
-    RETURN_NAMES = ("positive", "negative", "latent")
+            ],
+            outputs=[
+                io.Conditioning.Output(display_name="positive"),
+                io.Conditioning.Output(display_name="negative"),
+                io.Latent.Output(display_name="latent"),
+            ],
+        )
 
-    FUNCTION = "execute"
-
-    def execute(self, positive, negative, latent, noise_augmentation):
+    @classmethod
+    def execute(cls, positive, negative, latent, noise_augmentation) -> io.NodeOutput:
         latent = latent["samples"]
         positive = node_helpers.conditioning_set_values(positive, {"concat_latent_image": latent, "noise_augmentation": noise_augmentation})
         negative = node_helpers.conditioning_set_values(negative, {"concat_latent_image": latent, "noise_augmentation": noise_augmentation})
         out_latent = {}
         out_latent["samples"] = torch.zeros([latent.shape[0], 32, latent.shape[-3], latent.shape[-2], latent.shape[-1]], device=comfy.model_management.intermediate_device())
-        return (positive, negative, out_latent)
+        return io.NodeOutput(positive, negative, out_latent)
 
 
-NODE_CLASS_MAPPINGS = {
-    "CLIPTextEncodeHunyuanDiT": CLIPTextEncodeHunyuanDiT,
-    "TextEncodeHunyuanVideo_ImageToVideo": TextEncodeHunyuanVideo_ImageToVideo,
-    "EmptyHunyuanLatentVideo": EmptyHunyuanLatentVideo,
-    "HunyuanImageToVideo": HunyuanImageToVideo,
-    "EmptyHunyuanImageLatent": EmptyHunyuanImageLatent,
-    "HunyuanRefinerLatent": HunyuanRefinerLatent,
-}
+class HunyuanExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            CLIPTextEncodeHunyuanDiT,
+            TextEncodeHunyuanVideo_ImageToVideo,
+            EmptyHunyuanLatentVideo,
+            HunyuanImageToVideo,
+            EmptyHunyuanImageLatent,
+            HunyuanRefinerLatent,
+        ]
+
+
+async def comfy_entrypoint() -> HunyuanExtension:
+    return HunyuanExtension()

comfy_extras/nodes_latent.py

@@ -2,6 +2,8 @@ import comfy.utils
 import comfy_extras.nodes_post_processing
 import torch
 import nodes
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
 
 def reshape_latent_to(target_shape, latent, repeat_batch=True):
@@ -13,17 +15,23 @@ def reshape_latent_to(target_shape, latent, repeat_batch=True):
         return latent
 
 
-class LatentAdd:
+class LatentAdd(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples1": ("LATENT",), "samples2": ("LATENT",)}}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentAdd",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples1"),
+                io.Latent.Input("samples2"),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples1, samples2):
+    @classmethod
+    def execute(cls, samples1, samples2) -> io.NodeOutput:
         samples_out = samples1.copy()
 
         s1 = samples1["samples"]
@@ -31,19 +39,25 @@ class LatentAdd:
 
         s2 = reshape_latent_to(s1.shape, s2)
         samples_out["samples"] = s1 + s2
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentSubtract:
+class LatentSubtract(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples1": ("LATENT",), "samples2": ("LATENT",)}}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentSubtract",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples1"),
+                io.Latent.Input("samples2"),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples1, samples2):
+    @classmethod
+    def execute(cls, samples1, samples2) -> io.NodeOutput:
         samples_out = samples1.copy()
 
         s1 = samples1["samples"]
@@ -51,41 +65,49 @@ class LatentSubtract:
 
         s2 = reshape_latent_to(s1.shape, s2)
         samples_out["samples"] = s1 - s2
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentMultiply:
+class LatentMultiply(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples": ("LATENT",),
-                              "multiplier": ("FLOAT", {"default": 1.0, "min": -10.0, "max": 10.0, "step": 0.01}),
-                             }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentMultiply",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples"),
+                io.Float.Input("multiplier", default=1.0, min=-10.0, max=10.0, step=0.01),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples, multiplier):
+    @classmethod
+    def execute(cls, samples, multiplier) -> io.NodeOutput:
         samples_out = samples.copy()
 
         s1 = samples["samples"]
         samples_out["samples"] = s1 * multiplier
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentInterpolate:
+class LatentInterpolate(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples1": ("LATENT",),
-                              "samples2": ("LATENT",),
-                              "ratio": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.01}),
-                              }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentInterpolate",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples1"),
+                io.Latent.Input("samples2"),
+                io.Float.Input("ratio", default=1.0, min=0.0, max=1.0, step=0.01),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples1, samples2, ratio):
+    @classmethod
+    def execute(cls, samples1, samples2, ratio) -> io.NodeOutput:
         samples_out = samples1.copy()
 
         s1 = samples1["samples"]
@@ -104,19 +126,26 @@ class LatentInterpolate:
         st = torch.nan_to_num(t / mt)
 
         samples_out["samples"] = st * (m1 * ratio + m2 * (1.0 - ratio))
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentConcat:
+class LatentConcat(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples1": ("LATENT",), "samples2": ("LATENT",), "dim": (["x", "-x", "y", "-y", "t", "-t"], )}}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentConcat",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples1"),
+                io.Latent.Input("samples2"),
+                io.Combo.Input("dim", options=["x", "-x", "y", "-y", "t", "-t"]),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples1, samples2, dim):
+    @classmethod
+    def execute(cls, samples1, samples2, dim) -> io.NodeOutput:
         samples_out = samples1.copy()
 
         s1 = samples1["samples"]
@@ -136,22 +165,27 @@ class LatentConcat:
             dim = -3
 
         samples_out["samples"] = torch.cat(c, dim=dim)
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentCut:
+class LatentCut(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"samples": ("LATENT",),
-                             "dim": (["x", "y", "t"], ),
-                             "index": ("INT", {"default": 0, "min": -nodes.MAX_RESOLUTION, "max": nodes.MAX_RESOLUTION, "step": 1}),
-                             "amount": ("INT", {"default": 1, "min": 1, "max": nodes.MAX_RESOLUTION, "step": 1})}}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentCut",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples"),
+                io.Combo.Input("dim", options=["x", "y", "t"]),
+                io.Int.Input("index", default=0, min=-nodes.MAX_RESOLUTION, max=nodes.MAX_RESOLUTION, step=1),
+                io.Int.Input("amount", default=1, min=1, max=nodes.MAX_RESOLUTION, step=1),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples, dim, index, amount):
+    @classmethod
+    def execute(cls, samples, dim, index, amount) -> io.NodeOutput:
         samples_out = samples.copy()
 
         s1 = samples["samples"]
@@ -171,19 +205,25 @@ class LatentCut:
             amount = min(-index, amount)
 
         samples_out["samples"] = torch.narrow(s1, dim, index, amount)
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentBatch:
+class LatentBatch(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples1": ("LATENT",), "samples2": ("LATENT",)}}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentBatch",
+            category="latent/batch",
+            inputs=[
+                io.Latent.Input("samples1"),
+                io.Latent.Input("samples2"),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "batch"
-
-    CATEGORY = "latent/batch"
-
-    def batch(self, samples1, samples2):
+    @classmethod
+    def execute(cls, samples1, samples2) -> io.NodeOutput:
         samples_out = samples1.copy()
         s1 = samples1["samples"]
         s2 = samples2["samples"]
@@ -192,20 +232,25 @@ class LatentBatch:
         s = torch.cat((s1, s2), dim=0)
         samples_out["samples"] = s
         samples_out["batch_index"] = samples1.get("batch_index", [x for x in range(0, s1.shape[0])]) + samples2.get("batch_index", [x for x in range(0, s2.shape[0])])
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentBatchSeedBehavior:
+class LatentBatchSeedBehavior(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples": ("LATENT",),
-                              "seed_behavior": (["random", "fixed"],{"default": "fixed"}),}}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentBatchSeedBehavior",
+            category="latent/advanced",
+            inputs=[
+                io.Latent.Input("samples"),
+                io.Combo.Input("seed_behavior", options=["random", "fixed"], default="fixed"),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced"
-
-    def op(self, samples, seed_behavior):
+    @classmethod
+    def execute(cls, samples, seed_behavior) -> io.NodeOutput:
         samples_out = samples.copy()
         latent = samples["samples"]
         if seed_behavior == "random":
@@ -215,41 +260,50 @@ class LatentBatchSeedBehavior:
             batch_number = samples_out.get("batch_index", [0])[0]
             samples_out["batch_index"] = [batch_number] * latent.shape[0]
 
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentApplyOperation:
+class LatentApplyOperation(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "samples": ("LATENT",),
-                             "operation": ("LATENT_OPERATION",),
-                             }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentApplyOperation",
+            category="latent/advanced/operations",
+            is_experimental=True,
+            inputs=[
+                io.Latent.Input("samples"),
+                io.LatentOperation.Input("operation"),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced/operations"
-    EXPERIMENTAL = True
-
-    def op(self, samples, operation):
+    @classmethod
+    def execute(cls, samples, operation) -> io.NodeOutput:
         samples_out = samples.copy()
 
         s1 = samples["samples"]
         samples_out["samples"] = operation(latent=s1)
-        return (samples_out,)
+        return io.NodeOutput(samples_out)
 
-class LatentApplyOperationCFG:
+class LatentApplyOperationCFG(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "model": ("MODEL",),
-                             "operation": ("LATENT_OPERATION",),
-                              }}
-    RETURN_TYPES = ("MODEL",)
-    FUNCTION = "patch"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentApplyOperationCFG",
+            category="latent/advanced/operations",
+            is_experimental=True,
+            inputs=[
+                io.Model.Input("model"),
+                io.LatentOperation.Input("operation"),
+            ],
+            outputs=[
+                io.Model.Output(),
+            ],
+        )
 
-    CATEGORY = "latent/advanced/operations"
-    EXPERIMENTAL = True
-
-    def patch(self, model, operation):
+    @classmethod
+    def execute(cls, model, operation) -> io.NodeOutput:
         m = model.clone()
 
         def pre_cfg_function(args):
@@ -261,21 +315,25 @@ class LatentApplyOperationCFG:
             return conds_out
 
         m.set_model_sampler_pre_cfg_function(pre_cfg_function)
-        return (m, )
+        return io.NodeOutput(m)
 
-class LatentOperationTonemapReinhard:
+class LatentOperationTonemapReinhard(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "multiplier": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 100.0, "step": 0.01}),
-                              }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentOperationTonemapReinhard",
+            category="latent/advanced/operations",
+            is_experimental=True,
+            inputs=[
+                io.Float.Input("multiplier", default=1.0, min=0.0, max=100.0, step=0.01),
+            ],
+            outputs=[
+                io.LatentOperation.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT_OPERATION",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced/operations"
-    EXPERIMENTAL = True
-
-    def op(self, multiplier):
+    @classmethod
+    def execute(cls, multiplier) -> io.NodeOutput:
         def tonemap_reinhard(latent, **kwargs):
             latent_vector_magnitude = (torch.linalg.vector_norm(latent, dim=(1)) + 0.0000000001)[:,None]
             normalized_latent = latent / latent_vector_magnitude
@@ -291,39 +349,27 @@ class LatentOperationTonemapReinhard:
             new_magnitude *= top
 
             return normalized_latent * new_magnitude
-        return (tonemap_reinhard,)
+        return io.NodeOutput(tonemap_reinhard)
 
-class LatentOperationSharpen:
+class LatentOperationSharpen(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-                "sharpen_radius": ("INT", {
-                    "default": 9,
-                    "min": 1,
-                    "max": 31,
-                    "step": 1
-                }),
-                "sigma": ("FLOAT", {
-                    "default": 1.0,
-                    "min": 0.1,
-                    "max": 10.0,
-                    "step": 0.1
-                }),
-                "alpha": ("FLOAT", {
-                    "default": 0.1,
-                    "min": 0.0,
-                    "max": 5.0,
-                    "step": 0.01
-                }),
-                              }}
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LatentOperationSharpen",
+            category="latent/advanced/operations",
+            is_experimental=True,
+            inputs=[
+                io.Int.Input("sharpen_radius", default=9, min=1, max=31, step=1),
+                io.Float.Input("sigma", default=1.0, min=0.1, max=10.0, step=0.1),
+                io.Float.Input("alpha", default=0.1, min=0.0, max=5.0, step=0.01),
+            ],
+            outputs=[
+                io.LatentOperation.Output(),
+            ],
+        )
 
-    RETURN_TYPES = ("LATENT_OPERATION",)
-    FUNCTION = "op"
-
-    CATEGORY = "latent/advanced/operations"
-    EXPERIMENTAL = True
-
-    def op(self, sharpen_radius, sigma, alpha):
+    @classmethod
+    def execute(cls, sharpen_radius, sigma, alpha) -> io.NodeOutput:
         def sharpen(latent, **kwargs):
             luminance = (torch.linalg.vector_norm(latent, dim=(1)) + 1e-6)[:,None]
             normalized_latent = latent / luminance
@@ -340,19 +386,27 @@ class LatentOperationSharpen:
             sharpened = torch.nn.functional.conv2d(padded_image, kernel.repeat(channels, 1, 1).unsqueeze(1), padding=kernel_size // 2, groups=channels)[:,:,sharpen_radius:-sharpen_radius, sharpen_radius:-sharpen_radius]
 
             return luminance * sharpened
-        return (sharpen,)
+        return io.NodeOutput(sharpen)
 
-NODE_CLASS_MAPPINGS = {
-    "LatentAdd": LatentAdd,
-    "LatentSubtract": LatentSubtract,
-    "LatentMultiply": LatentMultiply,
-    "LatentInterpolate": LatentInterpolate,
-    "LatentConcat": LatentConcat,
-    "LatentCut": LatentCut,
-    "LatentBatch": LatentBatch,
-    "LatentBatchSeedBehavior": LatentBatchSeedBehavior,
-    "LatentApplyOperation": LatentApplyOperation,
-    "LatentApplyOperationCFG": LatentApplyOperationCFG,
-    "LatentOperationTonemapReinhard": LatentOperationTonemapReinhard,
-    "LatentOperationSharpen": LatentOperationSharpen,
-}
+
+class LatentExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            LatentAdd,
+            LatentSubtract,
+            LatentMultiply,
+            LatentInterpolate,
+            LatentConcat,
+            LatentCut,
+            LatentBatch,
+            LatentBatchSeedBehavior,
+            LatentApplyOperation,
+            LatentApplyOperationCFG,
+            LatentOperationTonemapReinhard,
+            LatentOperationSharpen,
+        ]
+
+
+async def comfy_entrypoint() -> LatentExtension:
+    return LatentExtension()

comfy_extras/nodes_lora_extract.py

@@ -5,6 +5,8 @@ import folder_paths
 import os
 import logging
 from enum import Enum
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
 CLAMP_QUANTILE = 0.99
 
@@ -71,32 +73,40 @@ def calc_lora_model(model_diff, rank, prefix_model, prefix_lora, output_sd, lora
             output_sd["{}{}.diff_b".format(prefix_lora, k[len(prefix_model):-5])] = sd[k].contiguous().half().cpu()
     return output_sd
 
-class LoraSave:
-    def __init__(self):
-        self.output_dir = folder_paths.get_output_directory()
+class LoraSave(io.ComfyNode):
+    @classmethod
+    def define_schema(cls):
+        return io.Schema(
+            node_id="LoraSave",
+            display_name="Extract and Save Lora",
+            category="_for_testing",
+            inputs=[
+                io.String.Input("filename_prefix", default="loras/ComfyUI_extracted_lora"),
+                io.Int.Input("rank", default=8, min=1, max=4096, step=1),
+                io.Combo.Input("lora_type", options=tuple(LORA_TYPES.keys())),
+                io.Boolean.Input("bias_diff", default=True),
+                io.Model.Input(
+                    "model_diff",
+                    tooltip="The ModelSubtract output to be converted to a lora.",
+                    optional=True,
+                ),
+                io.Clip.Input(
+                  "text_encoder_diff",
+                    tooltip="The CLIPSubtract output to be converted to a lora.",
+                    optional=True,
+                ),
+            ],
+            is_experimental=True,
+            is_output_node=True,
+        )
 
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"filename_prefix": ("STRING", {"default": "loras/ComfyUI_extracted_lora"}),
-                              "rank": ("INT", {"default": 8, "min": 1, "max": 4096, "step": 1}),
-                              "lora_type": (tuple(LORA_TYPES.keys()),),
-                              "bias_diff": ("BOOLEAN", {"default": True}),
-                            },
-                "optional": {"model_diff": ("MODEL", {"tooltip": "The ModelSubtract output to be converted to a lora."}),
-                             "text_encoder_diff": ("CLIP", {"tooltip": "The CLIPSubtract output to be converted to a lora."})},
-    }
-    RETURN_TYPES = ()
-    FUNCTION = "save"
-    OUTPUT_NODE = True
-
-    CATEGORY = "_for_testing"
-
-    def save(self, filename_prefix, rank, lora_type, bias_diff, model_diff=None, text_encoder_diff=None):
+    def execute(cls, filename_prefix, rank, lora_type, bias_diff, model_diff=None, text_encoder_diff=None) -> io.NodeOutput:
         if model_diff is None and text_encoder_diff is None:
-            return {}
+            return io.NodeOutput()
 
         lora_type = LORA_TYPES.get(lora_type)
-        full_output_folder, filename, counter, subfolder, filename_prefix = folder_paths.get_save_image_path(filename_prefix, self.output_dir)
+        full_output_folder, filename, counter, subfolder, filename_prefix = folder_paths.get_save_image_path(filename_prefix, folder_paths.get_output_directory())
 
         output_sd = {}
         if model_diff is not None:
@@ -108,12 +118,16 @@ class LoraSave:
         output_checkpoint = os.path.join(full_output_folder, output_checkpoint)
 
         comfy.utils.save_torch_file(output_sd, output_checkpoint, metadata=None)
-        return {}
+        return io.NodeOutput()
 
-NODE_CLASS_MAPPINGS = {
-    "LoraSave": LoraSave
-}
 
-NODE_DISPLAY_NAME_MAPPINGS = {
-    "LoraSave": "Extract and Save Lora"
-}
+class LoraSaveExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            LoraSave,
+        ]
+
+
+async def comfy_entrypoint() -> LoraSaveExtension:
+    return LoraSaveExtension()

comfy_extras/nodes_lt.py

@@ -34,6 +34,7 @@ class EmptyLTXVLatentVideo(io.ComfyNode):
         latent = torch.zeros([batch_size, 128, ((length - 1) // 8) + 1, height // 32, width // 32], device=comfy.model_management.intermediate_device())
         return io.NodeOutput({"samples": latent})
 
+    generate = execute  # TODO: remove
 
 class LTXVImgToVideo(io.ComfyNode):
     @classmethod
@@ -77,6 +78,8 @@ class LTXVImgToVideo(io.ComfyNode):
 
         return io.NodeOutput(positive, negative, {"samples": latent, "noise_mask": conditioning_latent_frames_mask})
 
+    generate = execute  # TODO: remove
+
 
 def conditioning_get_any_value(conditioning, key, default=None):
     for t in conditioning:
@@ -264,6 +267,8 @@ class LTXVAddGuide(io.ComfyNode):
 
         return io.NodeOutput(positive, negative, {"samples": latent_image, "noise_mask": noise_mask})
 
+    generate = execute  # TODO: remove
+
 
 class LTXVCropGuides(io.ComfyNode):
     @classmethod
@@ -300,6 +305,8 @@ class LTXVCropGuides(io.ComfyNode):
 
         return io.NodeOutput(positive, negative, {"samples": latent_image, "noise_mask": noise_mask})
 
+    crop = execute  # TODO: remove
+
 
 class LTXVConditioning(io.ComfyNode):
     @classmethod
@@ -498,6 +505,7 @@ class LTXVPreprocess(io.ComfyNode):
             output_images.append(preprocess(image[i], img_compression))
         return io.NodeOutput(torch.stack(output_images))
 
+    preprocess = execute  # TODO: remove
 
 class LtxvExtension(ComfyExtension):
     @override

comfy_extras/nodes_model_downscale.py

@@ -1,24 +1,33 @@
+from typing_extensions import override
 import comfy.utils
+from comfy_api.latest import ComfyExtension, io
 
-class PatchModelAddDownscale:
-    upscale_methods = ["bicubic", "nearest-exact", "bilinear", "area", "bislerp"]
+
+class PatchModelAddDownscale(io.ComfyNode):
+    UPSCALE_METHODS = ["bicubic", "nearest-exact", "bilinear", "area", "bislerp"]
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "model": ("MODEL",),
-                              "block_number": ("INT", {"default": 3, "min": 1, "max": 32, "step": 1}),
-                              "downscale_factor": ("FLOAT", {"default": 2.0, "min": 0.1, "max": 9.0, "step": 0.001}),
-                              "start_percent": ("FLOAT", {"default": 0.0, "min": 0.0, "max": 1.0, "step": 0.001}),
-                              "end_percent": ("FLOAT", {"default": 0.35, "min": 0.0, "max": 1.0, "step": 0.001}),
-                              "downscale_after_skip": ("BOOLEAN", {"default": True}),
-                              "downscale_method": (s.upscale_methods,),
-                              "upscale_method": (s.upscale_methods,),
-                              }}
-    RETURN_TYPES = ("MODEL",)
-    FUNCTION = "patch"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="PatchModelAddDownscale",
+            display_name="PatchModelAddDownscale (Kohya Deep Shrink)",
+            category="model_patches/unet",
+            inputs=[
+                io.Model.Input("model"),
+                io.Int.Input("block_number", default=3, min=1, max=32, step=1),
+                io.Float.Input("downscale_factor", default=2.0, min=0.1, max=9.0, step=0.001),
+                io.Float.Input("start_percent", default=0.0, min=0.0, max=1.0, step=0.001),
+                io.Float.Input("end_percent", default=0.35, min=0.0, max=1.0, step=0.001),
+                io.Boolean.Input("downscale_after_skip", default=True),
+                io.Combo.Input("downscale_method", options=cls.UPSCALE_METHODS),
+                io.Combo.Input("upscale_method", options=cls.UPSCALE_METHODS),
+            ],
+            outputs=[
+                io.Model.Output(),
+            ],
+        )
 
-    CATEGORY = "model_patches/unet"
-
-    def patch(self, model, block_number, downscale_factor, start_percent, end_percent, downscale_after_skip, downscale_method, upscale_method):
+    @classmethod
+    def execute(cls, model, block_number, downscale_factor, start_percent, end_percent, downscale_after_skip, downscale_method, upscale_method) -> io.NodeOutput:
         model_sampling = model.get_model_object("model_sampling")
         sigma_start = model_sampling.percent_to_sigma(start_percent)
         sigma_end = model_sampling.percent_to_sigma(end_percent)
@@ -41,13 +50,21 @@ class PatchModelAddDownscale:
         else:
             m.set_model_input_block_patch(input_block_patch)
         m.set_model_output_block_patch(output_block_patch)
-        return (m, )
+        return io.NodeOutput(m)
 
-NODE_CLASS_MAPPINGS = {
-    "PatchModelAddDownscale": PatchModelAddDownscale,
-}
 
 NODE_DISPLAY_NAME_MAPPINGS = {
     # Sampling
-    "PatchModelAddDownscale": "PatchModelAddDownscale (Kohya Deep Shrink)",
+    "PatchModelAddDownscale": "",
 }
+
+class ModelDownscaleExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            PatchModelAddDownscale,
+        ]
+
+
+async def comfy_entrypoint() -> ModelDownscaleExtension:
+    return ModelDownscaleExtension()

comfy_extras/nodes_preview_any.py

@@ -25,7 +25,7 @@ class PreviewAny():
             value = str(source)
         elif source is not None:
             try:
-                value = json.dumps(source)
+                value = json.dumps(source, indent=4)
             except Exception:
                 try:
                     value = str(source)

comfy_extras/nodes_sd3.py

@@ -3,64 +3,83 @@ import comfy.sd
 import comfy.model_management
 import nodes
 import torch
-import comfy_extras.nodes_slg
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
+from comfy_extras.nodes_slg import SkipLayerGuidanceDiT
 
 
-class TripleCLIPLoader:
+class TripleCLIPLoader(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "clip_name1": (folder_paths.get_filename_list("text_encoders"), ), "clip_name2": (folder_paths.get_filename_list("text_encoders"), ), "clip_name3": (folder_paths.get_filename_list("text_encoders"), )
-                             }}
-    RETURN_TYPES = ("CLIP",)
-    FUNCTION = "load_clip"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="TripleCLIPLoader",
+            category="advanced/loaders",
+            description="[Recipes]\n\nsd3: clip-l, clip-g, t5",
+            inputs=[
+                io.Combo.Input("clip_name1", options=folder_paths.get_filename_list("text_encoders")),
+                io.Combo.Input("clip_name2", options=folder_paths.get_filename_list("text_encoders")),
+                io.Combo.Input("clip_name3", options=folder_paths.get_filename_list("text_encoders")),
+            ],
+            outputs=[
+                io.Clip.Output(),
+            ],
+        )
 
-    CATEGORY = "advanced/loaders"
-
-    DESCRIPTION = "[Recipes]\n\nsd3: clip-l, clip-g, t5"
-
-    def load_clip(self, clip_name1, clip_name2, clip_name3):
+    @classmethod
+    def execute(cls, clip_name1, clip_name2, clip_name3) -> io.NodeOutput:
         clip_path1 = folder_paths.get_full_path_or_raise("text_encoders", clip_name1)
         clip_path2 = folder_paths.get_full_path_or_raise("text_encoders", clip_name2)
         clip_path3 = folder_paths.get_full_path_or_raise("text_encoders", clip_name3)
         clip = comfy.sd.load_clip(ckpt_paths=[clip_path1, clip_path2, clip_path3], embedding_directory=folder_paths.get_folder_paths("embeddings"))
-        return (clip,)
+        return io.NodeOutput(clip)
+
+    load_clip = execute  # TODO: remove
 
 
-class EmptySD3LatentImage:
-    def __init__(self):
-        self.device = comfy.model_management.intermediate_device()
+class EmptySD3LatentImage(io.ComfyNode):
+    @classmethod
+    def define_schema(cls):
+        return io.Schema(
+            node_id="EmptySD3LatentImage",
+            category="latent/sd3",
+            inputs=[
+                io.Int.Input("width", default=1024, min=16, max=nodes.MAX_RESOLUTION, step=16),
+                io.Int.Input("height", default=1024, min=16, max=nodes.MAX_RESOLUTION, step=16),
+                io.Int.Input("batch_size", default=1, min=1, max=4096),
+            ],
+            outputs=[
+                io.Latent.Output(),
+            ],
+        )
 
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "width": ("INT", {"default": 1024, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 16}),
-                              "height": ("INT", {"default": 1024, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 16}),
-                              "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096})}}
-    RETURN_TYPES = ("LATENT",)
-    FUNCTION = "generate"
+    def execute(cls, width, height, batch_size=1) -> io.NodeOutput:
+        latent = torch.zeros([batch_size, 16, height // 8, width // 8], device=comfy.model_management.intermediate_device())
+        return io.NodeOutput({"samples":latent})
 
-    CATEGORY = "latent/sd3"
-
-    def generate(self, width, height, batch_size=1):
-        latent = torch.zeros([batch_size, 16, height // 8, width // 8], device=self.device)
-        return ({"samples":latent}, )
+    generate = execute  # TODO: remove
 
 
-class CLIPTextEncodeSD3:
+class CLIPTextEncodeSD3(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {
-            "clip": ("CLIP", ),
-            "clip_l": ("STRING", {"multiline": True, "dynamicPrompts": True}),
-            "clip_g": ("STRING", {"multiline": True, "dynamicPrompts": True}),
-            "t5xxl": ("STRING", {"multiline": True, "dynamicPrompts": True}),
-            "empty_padding": (["none", "empty_prompt"], )
-            }}
-    RETURN_TYPES = ("CONDITIONING",)
-    FUNCTION = "encode"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="CLIPTextEncodeSD3",
+            category="advanced/conditioning",
+            inputs=[
+                io.Clip.Input("clip"),
+                io.String.Input("clip_l", multiline=True, dynamic_prompts=True),
+                io.String.Input("clip_g", multiline=True, dynamic_prompts=True),
+                io.String.Input("t5xxl", multiline=True, dynamic_prompts=True),
+                io.Combo.Input("empty_padding", options=["none", "empty_prompt"]),
+            ],
+            outputs=[
+                io.Conditioning.Output(),
+            ],
+        )
 
-    CATEGORY = "advanced/conditioning"
-
-    def encode(self, clip, clip_l, clip_g, t5xxl, empty_padding):
+    @classmethod
+    def execute(cls, clip, clip_l, clip_g, t5xxl, empty_padding) -> io.NodeOutput:
         no_padding = empty_padding == "none"
 
         tokens = clip.tokenize(clip_g)
@@ -82,57 +101,112 @@ class CLIPTextEncodeSD3:
                 tokens["l"] += empty["l"]
             while len(tokens["l"]) > len(tokens["g"]):
                 tokens["g"] += empty["g"]
-        return (clip.encode_from_tokens_scheduled(tokens), )
+        return io.NodeOutput(clip.encode_from_tokens_scheduled(tokens))
+
+    encode = execute  # TODO: remove
 
 
-class ControlNetApplySD3(nodes.ControlNetApplyAdvanced):
+class ControlNetApplySD3(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"positive": ("CONDITIONING", ),
-                             "negative": ("CONDITIONING", ),
-                             "control_net": ("CONTROL_NET", ),
-                             "vae": ("VAE", ),
-                             "image": ("IMAGE", ),
-                             "strength": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 10.0, "step": 0.01}),
-                             "start_percent": ("FLOAT", {"default": 0.0, "min": 0.0, "max": 1.0, "step": 0.001}),
-                             "end_percent": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.001})
-                             }}
-    CATEGORY = "conditioning/controlnet"
-    DEPRECATED = True
+    def define_schema(cls) -> io.Schema:
+        return io.Schema(
+            node_id="ControlNetApplySD3",
+            display_name="Apply Controlnet with VAE",
+            category="conditioning/controlnet",
+            inputs=[
+                io.Conditioning.Input("positive"),
+                io.Conditioning.Input("negative"),
+                io.ControlNet.Input("control_net"),
+                io.Vae.Input("vae"),
+                io.Image.Input("image"),
+                io.Float.Input("strength", default=1.0, min=0.0, max=10.0, step=0.01),
+                io.Float.Input("start_percent", default=0.0, min=0.0, max=1.0, step=0.001),
+                io.Float.Input("end_percent", default=1.0, min=0.0, max=1.0, step=0.001),
+            ],
+            outputs=[
+                io.Conditioning.Output(display_name="positive"),
+                io.Conditioning.Output(display_name="negative"),
+            ],
+            is_deprecated=True,
+        )
+
+    @classmethod
+    def execute(cls, positive, negative, control_net, image, strength, start_percent, end_percent, vae=None) -> io.NodeOutput:
+        if strength == 0:
+            return io.NodeOutput(positive, negative)
+
+        control_hint = image.movedim(-1, 1)
+        cnets = {}
+
+        out = []
+        for conditioning in [positive, negative]:
+            c = []
+            for t in conditioning:
+                d = t[1].copy()
+
+                prev_cnet = d.get('control', None)
+                if prev_cnet in cnets:
+                    c_net = cnets[prev_cnet]
+                else:
+                    c_net = control_net.copy().set_cond_hint(control_hint, strength, (start_percent, end_percent),
+                                                             vae=vae, extra_concat=[])
+                    c_net.set_previous_controlnet(prev_cnet)
+                    cnets[prev_cnet] = c_net
+
+                d['control'] = c_net
+                d['control_apply_to_uncond'] = False
+                n = [t[0], d]
+                c.append(n)
+            out.append(c)
+        return io.NodeOutput(out[0], out[1])
+
+    apply_controlnet = execute  # TODO: remove
 
 
-class SkipLayerGuidanceSD3(comfy_extras.nodes_slg.SkipLayerGuidanceDiT):
+class SkipLayerGuidanceSD3(io.ComfyNode):
     '''
     Enhance guidance towards detailed dtructure by having another set of CFG negative with skipped layers.
     Inspired by Perturbed Attention Guidance (https://arxiv.org/abs/2403.17377)
     Experimental implementation by Dango233@StabilityAI.
     '''
+
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"model": ("MODEL", ),
-                             "layers": ("STRING", {"default": "7, 8, 9", "multiline": False}),
-                             "scale": ("FLOAT", {"default": 3.0, "min": 0.0, "max": 10.0, "step": 0.1}),
-                             "start_percent": ("FLOAT", {"default": 0.01, "min": 0.0, "max": 1.0, "step": 0.001}),
-                             "end_percent": ("FLOAT", {"default": 0.15, "min": 0.0, "max": 1.0, "step": 0.001})
-                                }}
-    RETURN_TYPES = ("MODEL",)
-    FUNCTION = "skip_guidance_sd3"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="SkipLayerGuidanceSD3",
+            category="advanced/guidance",
+            description="Generic version of SkipLayerGuidance node that can be used on every DiT model.",
+            inputs=[
+                io.Model.Input("model"),
+                io.String.Input("layers", default="7, 8, 9", multiline=False),
+                io.Float.Input("scale", default=3.0, min=0.0, max=10.0, step=0.1),
+                io.Float.Input("start_percent", default=0.01, min=0.0, max=1.0, step=0.001),
+                io.Float.Input("end_percent", default=0.15, min=0.0, max=1.0, step=0.001),
+            ],
+            outputs=[
+                io.Model.Output(),
+            ],
+            is_experimental=True,
+        )
 
-    CATEGORY = "advanced/guidance"
+    @classmethod
+    def execute(cls, model, layers, scale, start_percent, end_percent) -> io.NodeOutput:
+        return SkipLayerGuidanceDiT().execute(model=model, scale=scale, start_percent=start_percent, end_percent=end_percent, double_layers=layers)
 
-    def skip_guidance_sd3(self, model, layers, scale, start_percent, end_percent):
-        return self.skip_guidance(model=model, scale=scale, start_percent=start_percent, end_percent=end_percent, double_layers=layers)
+    skip_guidance_sd3 = execute  # TODO: remove
 
 
-NODE_CLASS_MAPPINGS = {
-    "TripleCLIPLoader": TripleCLIPLoader,
-    "EmptySD3LatentImage": EmptySD3LatentImage,
-    "CLIPTextEncodeSD3": CLIPTextEncodeSD3,
-    "ControlNetApplySD3": ControlNetApplySD3,
-    "SkipLayerGuidanceSD3": SkipLayerGuidanceSD3,
-}
+class SD3Extension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            TripleCLIPLoader,
+            EmptySD3LatentImage,
+            CLIPTextEncodeSD3,
+            ControlNetApplySD3,
+            SkipLayerGuidanceSD3,
+        ]
 
-NODE_DISPLAY_NAME_MAPPINGS = {
-    # Sampling
-    "ControlNetApplySD3": "Apply Controlnet with VAE",
-}
+
+async def comfy_entrypoint() -> SD3Extension:
+    return SD3Extension()

comfy_extras/nodes_slg.py

@@ -1,33 +1,40 @@
 import comfy.model_patcher
 import comfy.samplers
 import re
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
 
-class SkipLayerGuidanceDiT:
+class SkipLayerGuidanceDiT(io.ComfyNode):
     '''
     Enhance guidance towards detailed dtructure by having another set of CFG negative with skipped layers.
     Inspired by Perturbed Attention Guidance (https://arxiv.org/abs/2403.17377)
     Original experimental implementation for SD3 by Dango233@StabilityAI.
     '''
+
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"model": ("MODEL", ),
-                             "double_layers": ("STRING", {"default": "7, 8, 9", "multiline": False}),
-                             "single_layers": ("STRING", {"default": "7, 8, 9", "multiline": False}),
-                             "scale": ("FLOAT", {"default": 3.0, "min": 0.0, "max": 10.0, "step": 0.1}),
-                             "start_percent": ("FLOAT", {"default": 0.01, "min": 0.0, "max": 1.0, "step": 0.001}),
-                             "end_percent": ("FLOAT", {"default": 0.15, "min": 0.0, "max": 1.0, "step": 0.001}),
-                             "rescaling_scale": ("FLOAT", {"default": 0.0, "min": 0.0, "max": 10.0, "step": 0.01}),
-                                }}
-    RETURN_TYPES = ("MODEL",)
-    FUNCTION = "skip_guidance"
-    EXPERIMENTAL = True
+    def define_schema(cls):
+        return io.Schema(
+            node_id="SkipLayerGuidanceDiT",
+            category="advanced/guidance",
+            description="Generic version of SkipLayerGuidance node that can be used on every DiT model.",
+            is_experimental=True,
+            inputs=[
+                io.Model.Input("model"),
+                io.String.Input("double_layers", default="7, 8, 9"),
+                io.String.Input("single_layers", default="7, 8, 9"),
+                io.Float.Input("scale", default=3.0, min=0.0, max=10.0, step=0.1),
+                io.Float.Input("start_percent", default=0.01, min=0.0, max=1.0, step=0.001),
+                io.Float.Input("end_percent", default=0.15, min=0.0, max=1.0, step=0.001),
+                io.Float.Input("rescaling_scale", default=0.0, min=0.0, max=10.0, step=0.01),
+            ],
+            outputs=[
+                io.Model.Output(),
+            ],
+        )
 
-    DESCRIPTION = "Generic version of SkipLayerGuidance node that can be used on every DiT model."
-
-    CATEGORY = "advanced/guidance"
-
-    def skip_guidance(self, model, scale, start_percent, end_percent, double_layers="", single_layers="", rescaling_scale=0):
+    @classmethod
+    def execute(cls, model, scale, start_percent, end_percent, double_layers="", single_layers="", rescaling_scale=0) -> io.NodeOutput:
         # check if layer is comma separated integers
         def skip(args, extra_args):
             return args
@@ -43,7 +50,7 @@ class SkipLayerGuidanceDiT:
         single_layers = [int(i) for i in single_layers]
 
         if len(double_layers) == 0 and len(single_layers) == 0:
-            return (model, )
+            return io.NodeOutput(model)
 
         def post_cfg_function(args):
             model = args["model"]
@@ -76,29 +83,36 @@ class SkipLayerGuidanceDiT:
         m = model.clone()
         m.set_model_sampler_post_cfg_function(post_cfg_function)
 
-        return (m, )
+        return io.NodeOutput(m)
 
-class SkipLayerGuidanceDiTSimple:
+    skip_guidance = execute  # TODO: remove
+
+
+class SkipLayerGuidanceDiTSimple(io.ComfyNode):
     '''
     Simple version of the SkipLayerGuidanceDiT node that only modifies the uncond pass.
     '''
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": {"model": ("MODEL", ),
-                             "double_layers": ("STRING", {"default": "7, 8, 9", "multiline": False}),
-                             "single_layers": ("STRING", {"default": "7, 8, 9", "multiline": False}),
-                             "start_percent": ("FLOAT", {"default": 0.0, "min": 0.0, "max": 1.0, "step": 0.001}),
-                             "end_percent": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.001}),
-                                }}
-    RETURN_TYPES = ("MODEL",)
-    FUNCTION = "skip_guidance"
-    EXPERIMENTAL = True
+    def define_schema(cls):
+        return io.Schema(
+            node_id="SkipLayerGuidanceDiTSimple",
+            category="advanced/guidance",
+            description="Simple version of the SkipLayerGuidanceDiT node that only modifies the uncond pass.",
+            is_experimental=True,
+            inputs=[
+                io.Model.Input("model"),
+                io.String.Input("double_layers", default="7, 8, 9"),
+                io.String.Input("single_layers", default="7, 8, 9"),
+                io.Float.Input("start_percent", default=0.0, min=0.0, max=1.0, step=0.001),
+                io.Float.Input("end_percent", default=1.0, min=0.0, max=1.0, step=0.001),
+            ],
+            outputs=[
+                io.Model.Output(),
+            ],
+        )
 
-    DESCRIPTION = "Simple version of the SkipLayerGuidanceDiT node that only modifies the uncond pass."
-
-    CATEGORY = "advanced/guidance"
-
-    def skip_guidance(self, model, start_percent, end_percent, double_layers="", single_layers=""):
+    @classmethod
+    def execute(cls, model, start_percent, end_percent, double_layers="", single_layers="") -> io.NodeOutput:
         def skip(args, extra_args):
             return args
 
@@ -113,7 +127,7 @@ class SkipLayerGuidanceDiTSimple:
         single_layers = [int(i) for i in single_layers]
 
         if len(double_layers) == 0 and len(single_layers) == 0:
-            return (model, )
+            return io.NodeOutput(model)
 
         def calc_cond_batch_function(args):
             x = args["input"]
@@ -144,9 +158,19 @@ class SkipLayerGuidanceDiTSimple:
         m = model.clone()
         m.set_model_sampler_calc_cond_batch_function(calc_cond_batch_function)
 
-        return (m, )
+        return io.NodeOutput(m)
 
-NODE_CLASS_MAPPINGS = {
-    "SkipLayerGuidanceDiT": SkipLayerGuidanceDiT,
-    "SkipLayerGuidanceDiTSimple": SkipLayerGuidanceDiTSimple,
-}
+    skip_guidance = execute  # TODO: remove
+
+
+class SkipLayerGuidanceExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            SkipLayerGuidanceDiT,
+            SkipLayerGuidanceDiTSimple,
+        ]
+
+
+async def comfy_entrypoint() -> SkipLayerGuidanceExtension:
+    return SkipLayerGuidanceExtension()

comfy_extras/nodes_upscale_model.py

@@ -4,6 +4,8 @@ from comfy import model_management
 import torch
 import comfy.utils
 import folder_paths
+from typing_extensions import override
+from comfy_api.latest import ComfyExtension, io
 
 try:
     from spandrel_extra_arches import EXTRA_REGISTRY
@@ -13,17 +15,23 @@ try:
 except:
     pass
 
-class UpscaleModelLoader:
+class UpscaleModelLoader(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "model_name": (folder_paths.get_filename_list("upscale_models"), ),
-                             }}
-    RETURN_TYPES = ("UPSCALE_MODEL",)
-    FUNCTION = "load_model"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="UpscaleModelLoader",
+            display_name="Load Upscale Model",
+            category="loaders",
+            inputs=[
+                io.Combo.Input("model_name", options=folder_paths.get_filename_list("upscale_models")),
+            ],
+            outputs=[
+                io.UpscaleModel.Output(),
+            ],
+        )
 
-    CATEGORY = "loaders"
-
-    def load_model(self, model_name):
+    @classmethod
+    def execute(cls, model_name) -> io.NodeOutput:
         model_path = folder_paths.get_full_path_or_raise("upscale_models", model_name)
         sd = comfy.utils.load_torch_file(model_path, safe_load=True)
         if "module.layers.0.residual_group.blocks.0.norm1.weight" in sd:
@@ -33,21 +41,29 @@ class UpscaleModelLoader:
         if not isinstance(out, ImageModelDescriptor):
             raise Exception("Upscale model must be a single-image model.")
 
-        return (out, )
+        return io.NodeOutput(out)
+
+    load_model = execute  # TODO: remove
 
 
-class ImageUpscaleWithModel:
+class ImageUpscaleWithModel(io.ComfyNode):
     @classmethod
-    def INPUT_TYPES(s):
-        return {"required": { "upscale_model": ("UPSCALE_MODEL",),
-                              "image": ("IMAGE",),
-                              }}
-    RETURN_TYPES = ("IMAGE",)
-    FUNCTION = "upscale"
+    def define_schema(cls):
+        return io.Schema(
+            node_id="ImageUpscaleWithModel",
+            display_name="Upscale Image (using Model)",
+            category="image/upscaling",
+            inputs=[
+                io.UpscaleModel.Input("upscale_model"),
+                io.Image.Input("image"),
+            ],
+            outputs=[
+                io.Image.Output(),
+            ],
+        )
 
-    CATEGORY = "image/upscaling"
-
-    def upscale(self, upscale_model, image):
+    @classmethod
+    def execute(cls, upscale_model, image) -> io.NodeOutput:
         device = model_management.get_torch_device()
 
         memory_required = model_management.module_size(upscale_model.model)
@@ -75,9 +91,19 @@ class ImageUpscaleWithModel:
 
         upscale_model.to("cpu")
         s = torch.clamp(s.movedim(-3,-1), min=0, max=1.0)
-        return (s,)
+        return io.NodeOutput(s)
 
-NODE_CLASS_MAPPINGS = {
-    "UpscaleModelLoader": UpscaleModelLoader,
-    "ImageUpscaleWithModel": ImageUpscaleWithModel
-}
+    upscale = execute  # TODO: remove
+
+
+class UpscaleModelExtension(ComfyExtension):
+    @override
+    async def get_node_list(self) -> list[type[io.ComfyNode]]:
+        return [
+            UpscaleModelLoader,
+            ImageUpscaleWithModel,
+        ]
+
+
+async def comfy_entrypoint() -> UpscaleModelExtension:
+    return UpscaleModelExtension()

comfyui_version.py

@@ -1,3 +1,3 @@
 # This file is automatically generated by the build process when version is
 # updated in pyproject.toml.
-__version__ = "0.3.63"
+__version__ = "0.3.66"

cuda_malloc.py

@@ -1,6 +1,6 @@
 import os
 import importlib.util
-from comfy.cli_args import args
+from comfy.cli_args import args, PerformanceFeature
 import subprocess
 
 #Can't use pytorch to get the GPU names because the cuda malloc has to be set before the first import.
@@ -75,8 +75,9 @@ if not args.cuda_malloc:
                 spec.loader.exec_module(module)
                 version = module.__version__
 
-        if int(version[0]) >= 2 and "+cu" in version: #enable by default for torch version 2.0 and up only on cuda torch
-            args.cuda_malloc = cuda_malloc_supported()
+        if int(version[0]) >= 2 and "+cu" in version:  # enable by default for torch version 2.0 and up only on cuda torch
+            if PerformanceFeature.AutoTune not in args.fast:  # Autotune has issues with cuda malloc
+                args.cuda_malloc = cuda_malloc_supported()
     except:
         pass
 

extra_model_paths.yaml.example

@@ -1,25 +1,5 @@
 #Rename this to extra_model_paths.yaml and ComfyUI will load it
 
-
-#config for a1111 ui
-#all you have to do is change the base_path to where yours is installed
-a111:
-    base_path: path/to/stable-diffusion-webui/
-
-    checkpoints: models/Stable-diffusion
-    configs: models/Stable-diffusion
-    vae: models/VAE
-    loras: |
-         models/Lora
-         models/LyCORIS
-    upscale_models: |
-                  models/ESRGAN
-                  models/RealESRGAN
-                  models/SwinIR
-    embeddings: embeddings
-    hypernetworks: models/hypernetworks
-    controlnet: models/ControlNet
-
 #config for comfyui
 #your base path should be either an existing comfy install or a central folder where you store all of your models, loras, etc.
 
@@ -28,7 +8,9 @@ a111:
 #     # You can use is_default to mark that these folders should be listed first, and used as the default dirs for eg downloads
 #     #is_default: true
 #     checkpoints: models/checkpoints/
-#     clip: models/clip/
+#     text_encoders: |
+#          models/text_encoders/
+#          models/clip/  # legacy location still supported
 #     clip_vision: models/clip_vision/
 #     configs: models/configs/
 #     controlnet: models/controlnet/
@@ -39,6 +21,32 @@ a111:
 #     loras: models/loras/
 #     upscale_models: models/upscale_models/
 #     vae: models/vae/
+#     audio_encoders: models/audio_encoders/
+#     model_patches: models/model_patches/
+
+
+#config for a1111 ui
+#all you have to do is uncomment this (remove the #) and change the base_path to where yours is installed
+
+#a111:
+#     base_path: path/to/stable-diffusion-webui/
+#     checkpoints: models/Stable-diffusion
+#     configs: models/Stable-diffusion
+#     vae: models/VAE
+#     loras: |
+#          models/Lora
+#          models/LyCORIS
+#     upscale_models: |
+#                   models/ESRGAN
+#                   models/RealESRGAN
+#                   models/SwinIR
+#     embeddings: embeddings
+#     hypernetworks: models/hypernetworks
+#     controlnet: models/ControlNet
+
+
+# For a full list of supported keys (style_models, vae_approx, hypernetworks, photomaker,
+# model_patches, audio_encoders, classifiers, etc.) see folder_paths.py.
 
 #other_ui:
 #    base_path: path/to/ui

nodes.py

@@ -2027,7 +2027,6 @@ NODE_DISPLAY_NAME_MAPPINGS = {
     "DiffControlNetLoader": "Load ControlNet Model (diff)",
     "StyleModelLoader": "Load Style Model",
     "CLIPVisionLoader": "Load CLIP Vision",
-    "UpscaleModelLoader": "Load Upscale Model",
     "UNETLoader": "Load Diffusion Model",
     # Conditioning
     "CLIPVisionEncode": "CLIP Vision Encode",
@@ -2065,7 +2064,6 @@ NODE_DISPLAY_NAME_MAPPINGS = {
     "LoadImageOutput": "Load Image (from Outputs)",
     "ImageScale": "Upscale Image",
     "ImageScaleBy": "Upscale Image By",
-    "ImageUpscaleWithModel": "Upscale Image (using Model)",
     "ImageInvert": "Invert Image",
     "ImagePadForOutpaint": "Pad Image for Outpainting",
     "ImageBatch": "Batch Images",
@@ -2357,6 +2355,7 @@ async def init_builtin_api_nodes():
         "nodes_stability.py",
         "nodes_pika.py",
         "nodes_runway.py",
+        "nodes_sora.py",
         "nodes_tripo.py",
         "nodes_moonvalley.py",
         "nodes_rodin.py",

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "ComfyUI"
-version = "0.3.63"
+version = "0.3.66"
 readme = "README.md"
 license = { file = "LICENSE" }
 requires-python = ">=3.9"
@@ -57,18 +57,13 @@ messages_control.disable = [
   "redefined-builtin",
   "unnecessary-lambda",
   "dangerous-default-value",
+  "invalid-overridden-method",
   # next warnings should be fixed in future
   "bad-classmethod-argument",  # Class method should have 'cls' as first argument
   "wrong-import-order",  # Standard imports should be placed before third party imports
-  "logging-fstring-interpolation", # Use lazy % formatting in logging functions
   "ungrouped-imports",
   "unnecessary-pass",
-  "unidiomatic-typecheck",
   "unnecessary-lambda-assignment",
   "no-else-return",
-  "no-else-raise",
-  "invalid-overridden-method",
   "unused-variable",
-  "pointless-string-statement",
-  "redefined-outer-name",
 ]

requirements.txt

@@ -1,6 +1,6 @@
-comfyui-frontend-package==1.27.7
-comfyui-workflow-templates==0.1.93
-comfyui-embedded-docs==0.2.6
+comfyui-frontend-package==1.28.7
+comfyui-workflow-templates==0.2.1
+comfyui-embedded-docs==0.3.0
 torch
 torchsde
 torchvision

server.py

@@ -48,6 +48,28 @@ async def send_socket_catch_exception(function, message):
     except (aiohttp.ClientError, aiohttp.ClientPayloadError, ConnectionResetError, BrokenPipeError, ConnectionError) as err:
         logging.warning("send error: {}".format(err))
 
+# Track deprecated paths that have been warned about to only warn once per file
+_deprecated_paths_warned = set()
+
+@web.middleware
+async def deprecation_warning(request: web.Request, handler):
+    """Middleware to warn about deprecated frontend API paths"""
+    path = request.path
+
+    if path.startswith("/scripts/ui") or path.startswith("/extensions/core/"):
+        # Only warn once per unique file path
+        if path not in _deprecated_paths_warned:
+            _deprecated_paths_warned.add(path)
+            logging.warning(
+                f"[DEPRECATION WARNING] Detected import of deprecated legacy API: {path}. "
+                f"This is likely caused by a custom node extension using outdated APIs. "
+                f"Please update your extensions or contact the extension author for an updated version."
+            )
+
+    response: web.Response = await handler(request)
+    return response
+
+
 @web.middleware
 async def compress_body(request: web.Request, handler):
     accept_encoding = request.headers.get("Accept-Encoding", "")
@@ -159,7 +181,7 @@ class PromptServer():
         self.client_session:Optional[aiohttp.ClientSession] = None
         self.number = 0
 
-        middlewares = [cache_control]
+        middlewares = [cache_control, deprecation_warning]
         if args.enable_compress_response_body:
             middlewares.append(compress_body)