[PR #1602] [CLOSED] feat: (pending closure) separate openai speech API provider #36350

New Issue

GiteaMirror · 2026-04-25T10:16:08-05:00

GiteaMirror commented

2026-04-25 10:16:08 -05:00

📋 Pull Request Information

Original PR: https://github.com/open-webui/open-webui/pull/1602
Author: @lee-b
Created: 4/18/2024
Status: ❌ Closed

Base: main ← Head: feature/hacky-openai-speech-api-support

📝 Commits (6)

fe01499 Ignore ollama-data dir
19cde7d Support a local openedai-speech server separate from ollama
2a90810 Ignore .vscode/
8719cf7 Properly join paths
25b34ac Ensure speech URL path part is joined relative to base
6c8dfa9 Rename OPENAI_BASE_URL to TTS_OPENAI_API_BASE_URL per #1602 and Error 429 (#1569)

📊 Changes

5 files changed (+69 additions, -6 deletions)

View changed files

📝 .gitignore (+5 -1)
📝 backend/apps/openai/main.py (+30 -4)
➕ docker-compose.customized.yaml (+7 -0)
➕ run-customized.sh (+25 -0)
📝 src/lib/apis/openai/index.ts (+2 -1)

📄 Description

WORK IN PROGRESS

FILING THIS PR MORE FOR AWARENESS/TRACKING RIGHT NOW - SUGGESTIONS WELCOME

In particular, I should:

merge the docker-customized stuff into the repo's own docker files and startup scripts
probably try to implement authenticated remote openai support (though I don't care for it personally) or just not have stub code in there (is_openedai_speech) that hints at future support, for now.

Description

Add support for a separate OpenAI-compatible speech API base URL.

This allows the use of Ollama or an OpenAI-compatible service such as Llama.CPP for text gen (which may be faster / easier for text generation), and another API provider like LocalAI or OpenedAI-Speech for the text-to-speech and speech to text functionality.

Changelog Entry

Added

Add support for a separate OpenAI-compatible speech API providers, using an OPENAI_AUDIO_BASE_URL environment variable.
docker-compose.customized.yaml which sets the new env vars
run_customized.sh which simplifies running openai with the above customized config

Fixed

Allows use of llama.cpp or another text-only OpenAI endpoint / model whilst using a separate/dedicated speech API provider

Changed

Supports a new OPENAI_AUDIO_BASE_URL, which will be used rather than OPENAI_BASE_URL for speech services, when set.

_{🔄 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/open-webui/pull/1602 **Author:** [@lee-b](https://github.com/lee-b) **Created:** 4/18/2024 **Status:** ❌ Closed **Base:** `main` ← **Head:** `feature/hacky-openai-speech-api-support` --- ### 📝 Commits (6) - [`fe01499`](https://github.com/open-webui/open-webui/commit/fe014999e91f1417d624500be3a37d08e5003ffb) Ignore ollama-data dir - [`19cde7d`](https://github.com/open-webui/open-webui/commit/19cde7dcea24732daa07bce11da0b22a7f280df1) Support a local openedai-speech server separate from ollama - [`2a90810`](https://github.com/open-webui/open-webui/commit/2a908109b85425253674a8623e9a85f4aeb6a054) Ignore .vscode/ - [`8719cf7`](https://github.com/open-webui/open-webui/commit/8719cf7e918993f8f1b14d9241e6069b5b4fdf1e) Properly join paths - [`25b34ac`](https://github.com/open-webui/open-webui/commit/25b34ac3647b9a667fd7712438ae6d9511954ccc) Ensure speech URL path part is joined relative to base - [`6c8dfa9`](https://github.com/open-webui/open-webui/commit/6c8dfa96fc7a9aca2a675895d8649050365b7eff) Rename OPENAI_BASE_URL to TTS_OPENAI_API_BASE_URL per #1602 and #1569 ### 📊 Changes **5 files changed** (+69 additions, -6 deletions) <details> <summary>View changed files</summary> 📝 `.gitignore` (+5 -1) 📝 `backend/apps/openai/main.py` (+30 -4) ➕ `docker-compose.customized.yaml` (+7 -0) ➕ `run-customized.sh` (+25 -0) 📝 `src/lib/apis/openai/index.ts` (+2 -1) </details> ### 📄 Description WORK IN PROGRESS FILING THIS PR MORE FOR AWARENESS/TRACKING RIGHT NOW - SUGGESTIONS WELCOME In particular, I should: - merge the docker-customized stuff into the repo's own docker files and startup scripts - probably try to implement authenticated remote openai support (though I don't care for it personally) or just not have stub code in there (is_openedai_speech) that hints at future support, for now. ## Description Add support for a separate OpenAI-compatible speech API base URL. This allows the use of Ollama or an OpenAI-compatible service such as Llama.CPP for text gen (which may be faster / easier for text generation), and another API provider like LocalAI or OpenedAI-Speech for the text-to-speech and speech to text functionality. --- ### Changelog Entry ### Added - Add support for a separate OpenAI-compatible speech API providers, using an OPENAI_AUDIO_BASE_URL environment variable. - docker-compose.customized.yaml which sets the new env vars - run_customized.sh which simplifies running openai with the above customized config ### Fixed - Allows use of llama.cpp or another text-only OpenAI endpoint / model whilst using a separate/dedicated speech API provider ### Changed - Supports a new OPENAI_AUDIO_BASE_URL, which will be used rather than OPENAI_BASE_URL for speech services, when set. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

GiteaMirror added the pull-request label 2026-04-25 10:16:08 -05:00

GiteaMirror closed this issue

2026-04-25 10:16:09 -05:00

Sign in to join this conversation.

Branches Tags

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: github-starred/open-webui#36350