[PR #1417] [MERGED] docs: explain self-hosted OpenAI-compatible speech-to-text #1382

New issue

Closed

opened 2026-10-18 14:19:17 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-18 14:19:17 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/open-webui/docs/pull/1417
Author: @Nathan-Roll1
Created: 9/18/2026
Status: ✅ Merged
Merged: 9/18/2026
Merged by: @Classic298

Base: main ← Head: codex/self-hosted-stt-guide

📝 Commits (1)

bce5cd1 docs: explain self-hosted OpenAI-compatible speech-to-text

📊 Changes

2 files changed (+98 additions, -0 deletions)

View changed files

➕ docs/features/chat-conversations/audio/speech-to-text/self-hosted-stt.md (+97 -0)
📝 docs/features/chat-conversations/audio/speech-to-text/stt-config.md (+1 -0)

📄 Description

Summary

Add a generic guide for connecting Open WebUI to a self-hosted OpenAI-compatible transcription endpoint. The worked example uses LocalAI's stock Whisper base model. It starts with the endpoint, key, and model settings, then covers native setup, container addressing, and common errors. The STT provider table links to the guide.

Disclosure: I founded Oruk. My earlier PR #1392 used our model and was closed because it was promotional and lacked that disclosure. This guide follows Classic298's invitation to document a generic setup with an unaffiliated model. It contains no Oruk model, service, links, or download-accounting material.

Maintainer feedback on #1392. That PR remains closed.

Checklist

I have reviewed the relevant documentation and matched the existing style.
This PR meets Open WebUI's contribution standards: it is accurate, relevant to users, narrowly scoped, maintainable, and not promotional content, advertising, lead generation, SEO placement, or a request to list a product, service, provider, integration, gateway, tool, or company primarily for visibility.
I understand that PRs that do not meet these standards may be closed without review and will not be merged. Repeated, low-quality, off-topic, promotional, or intentionally misleading submissions may result in the contributor being blocked from future participation in Open WebUI repositories.

Notes for reviewers

Validated with Open WebUI 0.11.3 and LocalAI 4.9.0 on macOS arm64. Installed the stock whisper-1 gallery entry with the documented commands and verified its file against the gallery checksum. Through Open WebUI's actual transcription endpoint, repeated English WAV and WebM/Opus recordings returned the expected transcript; an invalid extension returned 400, a valid request afterward succeeded, and LocalAI rejected an incorrect key with 401. Saved STT configuration also round-tripped. One- and five-second silence clips returned Whisper's [BLANK_AUDIO] marker. These are functional checks, not an accuracy or performance benchmark.

The documentation production build passed (426 pages). Desktop and mobile rendering were inspected. In Chromium, a fake microphone fed with real English audio produced a browser recording, the expected transcript appeared in the chat input, and the visible admin STT settings saved successfully. Container deployment and a physical microphone have not been tested. This is limited to one new guide and one provider-table link; it changes no application code or dependencies.

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

## 📋 Pull Request Information **Original PR:** https://github.com/open-webui/docs/pull/1417 **Author:** [@Nathan-Roll1](https://github.com/Nathan-Roll1) **Created:** 9/18/2026 **Status:** ✅ Merged **Merged:** 9/18/2026 **Merged by:** [@Classic298](https://github.com/Classic298) **Base:** `main` ← **Head:** `codex/self-hosted-stt-guide` --- ### 📝 Commits (1) - [`bce5cd1`](https://github.com/open-webui/docs/commit/bce5cd1fce4cbdb5e75b9e2657914e466513432d) docs: explain self-hosted OpenAI-compatible speech-to-text ### 📊 Changes **2 files changed** (+98 additions, -0 deletions) <details> <summary>View changed files</summary> ➕ `docs/features/chat-conversations/audio/speech-to-text/self-hosted-stt.md` (+97 -0) 📝 `docs/features/chat-conversations/audio/speech-to-text/stt-config.md` (+1 -0) </details> ### 📄 Description ## Summary Add a generic guide for connecting Open WebUI to a self-hosted OpenAI-compatible transcription endpoint. The worked example uses LocalAI's stock Whisper base model. It starts with the endpoint, key, and model settings, then covers native setup, container addressing, and common errors. The STT provider table links to the guide. Disclosure: I founded Oruk. My earlier PR #1392 used our model and was closed because it was promotional and lacked that disclosure. This guide follows Classic298's invitation to document a generic setup with an unaffiliated model. It contains no Oruk model, service, links, or download-accounting material. ## Related issue or discussion [Maintainer feedback on #1392](https://github.com/open-webui/docs/pull/1392#issuecomment-5727374942). That PR remains closed. ## Checklist - [x] I have reviewed the relevant documentation and matched the existing style. - [x] This PR meets Open WebUI's contribution standards: it is accurate, relevant to users, narrowly scoped, maintainable, and not promotional content, advertising, lead generation, SEO placement, or a request to list a product, service, provider, integration, gateway, tool, or company primarily for visibility. - [x] I understand that PRs that do not meet these standards may be closed without review and will not be merged. Repeated, low-quality, off-topic, promotional, or intentionally misleading submissions may result in the contributor being blocked from future participation in Open WebUI repositories. ## Notes for reviewers Validated with Open WebUI 0.11.3 and LocalAI 4.9.0 on macOS arm64. Installed the stock `whisper-1` gallery entry with the documented commands and verified its file against the gallery checksum. Through Open WebUI's actual transcription endpoint, repeated English WAV and WebM/Opus recordings returned the expected transcript; an invalid extension returned 400, a valid request afterward succeeded, and LocalAI rejected an incorrect key with 401. Saved STT configuration also round-tripped. One- and five-second silence clips returned Whisper's `[BLANK_AUDIO]` marker. These are functional checks, not an accuracy or performance benchmark. The documentation production build passed (426 pages). Desktop and mobile rendering were inspected. In Chromium, a fake microphone fed with real English audio produced a browser recording, the expected transcript appeared in the chat input, and the visible admin STT settings saved successfully. Container deployment and a physical microphone have not been tested. This is limited to one new guide and one provider-table link; it changes no application code or dependencies. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-10-18 14:19:17 +00:00