[PR #450] docs: add Gneol integration guide (EN + zh-CN)

## 📋 Pull Request Information **Original PR:** https://github.com/deepseek-ai/awesome-deepseek-agent/pull/450 **Author:** [@Elijahdanie](https://github.com/Elijahdanie) **Created:** 10/12/2026 **Status:** 🔄 Open **Base:** `main` ← **Head:** `add-gneol-guide` --- ### 📝 Commits (1) - [`cb71fd1`](https://github.com/deepseek-ai/awesome-deepseek-agent/commit/cb71fd16c4f832778ee1a3590c404fc00ccbf766) docs: add Gneol integration guide (EN + zh-CN) ### 📊 Changes **4 files changed** (+280 additions, -0 deletions) <details> <summary>View changed files</summary> 📝 `README.md` (+1 -0) 📝 `README.zh-CN.md` (+1 -0) ➕ `docs/gneol.md` (+139 -0) ➕ `docs/gneol.zh-CN.md` (+139 -0) </details> ### 📄 Description ## Summary Adds a bilingual integration guide for **Gneol** — a runtime for `.gneol` files. One plain-text file describes an AI agent (model, subagents, scheduled triggers, event listeners, tools, MCP servers) and Gneol runs it locally as a single process. - `docs/gneol.md` (English) - `docs/gneol.zh-CN.md` (简体中文) - README table entries added in **both** `README.md` and `README.zh-CN.md`, inserted alphabetically between **GitHub Copilot CLI** and **Hermes**. Gneol connects to DeepSeek through its built-in `deepseek` provider, which speaks the OpenAI-compatible Chat Completions API. ## Critical checks - [x] **Model naming** — uses `deepseek-v4-pro` / `deepseek-v4-flash` only. No V3 names (`deepseek-chat` / `deepseek-reasoner` / `deepseek-coder`). - [x] **1M context** — Gneol is OpenAI-compatible, so the guide sets `.maxTokens(384000)` and states in prose that DeepSeek V4 supports 1M context. `context_window` is not a Gneol config field, so it is called out in prose rather than faked in config. - [x] **Max thinking effort** — noted for `deepseek-v4-pro`. Gneol currently forwards `temperature` / `maxTokens` to the OpenAI-compatible endpoint; `reasoning_effort` is not yet exposed as a `.gneol` field, so the guide does not claim a config key that does not exist. - [x] **No pricing table** — nothing to verify. - [x] **Config fields verified** — the documented `.gneol` block was parsed with `gneol convert -f hello.gneol -t json`; it resolves `model_0` with `maxTokens: 384000`. Every directive shown (`.provider`, `.modelId`, `.apiKey`, `.temperature`, `.maxTokens`, `.env`, `subagent()`, `summarization()`) exists and is parsed. - [x] **No workarounds for upstream bugs** — none documented. ## Documentation - [x] English + Chinese in a single PR - [x] README entries in alphabetical order - [x] One tool per PR - [x] No images ## Notes `.maxTokens` is the max **output** tokens, so it is set to DeepSeek's `384000` output ceiling; the 1M input window is described in prose. Happy to adjust if you'd prefer different phrasing. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi commented

2026-10-26 09:49:32 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/deepseek-ai/awesome-deepseek-agent/pull/450
Author: @Elijahdanie
Created: 10/12/2026
Status: 🔄 Open

Base: main ← Head: add-gneol-guide

📝 Commits (1)

cb71fd1 docs: add Gneol integration guide (EN + zh-CN)

📊 Changes

4 files changed (+280 additions, -0 deletions)

View changed files

📝 README.md (+1 -0)
📝 README.zh-CN.md (+1 -0)
➕ docs/gneol.md (+139 -0)
➕ docs/gneol.zh-CN.md (+139 -0)

📄 Description

Summary

Adds a bilingual integration guide for Gneol — a runtime for .gneol files. One plain-text file describes an AI agent (model, subagents, scheduled triggers, event listeners, tools, MCP servers) and Gneol runs it locally as a single process.

docs/gneol.md (English)
docs/gneol.zh-CN.md (简体中文)
README table entries added in both README.md and README.zh-CN.md, inserted alphabetically between GitHub Copilot CLI and Hermes.

Gneol connects to DeepSeek through its built-in deepseek provider, which speaks the OpenAI-compatible Chat Completions API.

Critical checks

Model naming — uses deepseek-v4-pro / deepseek-v4-flash only. No V3 names (deepseek-chat / deepseek-reasoner / deepseek-coder).
1M context — Gneol is OpenAI-compatible, so the guide sets .maxTokens(384000) and states in prose that DeepSeek V4 supports 1M context. context_window is not a Gneol config field, so it is called out in prose rather than faked in config.
Max thinking effort — noted for deepseek-v4-pro. Gneol currently forwards temperature / maxTokens to the OpenAI-compatible endpoint; reasoning_effort is not yet exposed as a .gneol field, so the guide does not claim a config key that does not exist.
No pricing table — nothing to verify.
Config fields verified — the documented .gneol block was parsed with gneol convert -f hello.gneol -t json; it resolves model_0 with maxTokens: 384000. Every directive shown (.provider, .modelId, .apiKey, .temperature, .maxTokens, .env, subagent(), summarization()) exists and is parsed.
No workarounds for upstream bugs — none documented.

Documentation

English + Chinese in a single PR
README entries in alphabetical order
One tool per PR
No images

Notes

.maxTokens is the max output tokens, so it is set to DeepSeek's 384000 output ceiling; the 1M input window is described in prose. Happy to adjust if you'd prefer different phrasing.

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

bsakaguchi added the

pull-request

label

Rows
Columns

[PR #450] docs: add Gneol integration guide (EN + zh-CN) #446