[GH-ISSUE #17316] issue: Text-to-Speech broken after latest update when using local TTS provider via OpenAI engine #72842

New Issue

GiteaMirror · 2026-05-13T05:01:37-05:00

GiteaMirror commented

2026-05-13 05:01:37 -05:00

Originally created by @robertbeckey on GitHub (Sep 9, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/17316

Check Existing Issues

I have searched for any existing and/or related issues.
I have searched for any existing and/or related discussions.
I am using the latest version of Open WebUI.

Installation Method

Docker

Open WebUI Version

v0.6.27

Ollama Version (if applicable)

No response

Operating System

Windows 11

Browser (if applicable)

Brave

Confirmation

I have read and followed all instructions in README.md.
I am using the latest version of both Open WebUI and Ollama.
I have included the browser console logs.
I have included the Docker container logs.
I have provided every relevant configuration, setting, and environment variable used in my setup.
I have clearly listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
I have documented step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation. My steps:
Start with the initial platform/version/OS and dependencies used,
Specify exact install/launch/configure commands,
List URLs visited, user input (incl. example values/emails/passwords if needed),
Describe all options and toggles enabled or changed,
Include any files or environmental changes,
Identify the expected and actual result at each stage,
Ensure any reasonably skilled user can follow and hit the same issue.

Expected Behavior

In previous versions, text-to-speech generated correctly using the following options:
Text-to-Speech Engine: OpenAI
API base URL: http://host.docker.internal:8880/v1
API key: not-needed
TTS Voice: af_heart
TTS Model: kokoro
Response splitting: Punctuation

This would connect up to a kokoro-fastapi Docker container for local TTS and worked great.

Actual Behavior

Now, the Read Aloud button produces the following pop-up error many times in Open WebUI:
External: 404, message='Not Found', url='http://host.docker.internal:8880/audio/speech'

In the kokoro-fastapi container logs, the following error is posted:
"POST /audio/speech HTTP/1.1" 404 Not Found

Steps to Reproduce

Already described

Logs & Screenshots

Already described

Additional Information

No response

Originally created by @robertbeckey on GitHub (Sep 9, 2025). Original GitHub issue: https://github.com/open-webui/open-webui/issues/17316 ### Check Existing Issues - [x] I have searched for any existing and/or related issues. - [x] I have searched for any existing and/or related discussions. - [x] I am using the latest version of Open WebUI. ### Installation Method Docker ### Open WebUI Version v0.6.27 ### Ollama Version (if applicable) _No response_ ### Operating System Windows 11 ### Browser (if applicable) Brave ### Confirmation - [x] I have read and followed all instructions in `README.md`. - [x] I am using the latest version of **both** Open WebUI and Ollama. - [x] I have included the browser console logs. - [x] I have included the Docker container logs. - [x] I have **provided every relevant configuration, setting, and environment variable used in my setup.** - [x] I have clearly **listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup** (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc). - [x] I have documented **step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation**. My steps: - Start with the initial platform/version/OS and dependencies used, - Specify exact install/launch/configure commands, - List URLs visited, user input (incl. example values/emails/passwords if needed), - Describe all options and toggles enabled or changed, - Include any files or environmental changes, - Identify the expected and actual result at each stage, - Ensure any reasonably skilled user can follow and hit the same issue. ### Expected Behavior In previous versions, text-to-speech generated correctly using the following options: Text-to-Speech Engine: OpenAI API base URL: http://host.docker.internal:8880/v1 API key: not-needed TTS Voice: af_heart TTS Model: kokoro Response splitting: Punctuation This would connect up to a kokoro-fastapi Docker container for local TTS and worked great. ### Actual Behavior Now, the Read Aloud button produces the following pop-up error many times in Open WebUI: External: 404, message='Not Found', url='http://host.docker.internal:8880/audio/speech' In the kokoro-fastapi container logs, the following error is posted: "POST /audio/speech HTTP/1.1" 404 Not Found ### Steps to Reproduce Already described ### Logs & Screenshots Already described ### Additional Information _No response_

GiteaMirror added the bug label 2026-05-13 05:01:37 -05:00

GiteaMirror closed this issue

2026-05-13 05:01:38 -05:00

GiteaMirror commented

2026-05-13 05:01:39 -05:00

@98h398hrpohpoai commented on GitHub (Sep 9, 2025):

Same. PWA shows the api no longer passing the /v1 endpoint, but rather /audio/speech (without v1)

Edit: Here's the reference for it: https://github.com/open-webui/open-webui/pull/17061#issuecomment-3249144572

@98h398hrpohpoai commented on GitHub (Sep 9, 2025): Same. PWA shows the api no longer passing the /v1 endpoint, but rather /audio/speech (without v1) Edit: Here's the reference for it: https://github.com/open-webui/open-webui/pull/17061#issuecomment-3249144572

GiteaMirror commented

2026-05-13 05:01:41 -05:00

@98h398hrpohpoai commented on GitHub (Sep 9, 2025):

This quick fix worked for me (I'm dumb, don't assume this won't break other stuff--I have no idea):
Remove the leading slash from the relative path of the join statement outlined below.

url=urljoin(
    request.app.state.config.TTS_OPENAI_API_BASE_URL,
    "/audio/speech",

to

url=urljoin(
    request.app.state.config.TTS_OPENAI_API_BASE_URL,
    "audio/speech",

In line 342 of open_webui/routers/audio.py

Also ensure the base tts URL ends with a slash (because we can't use a leading slash in the join function) e.g. https://tts.localhost/v1/

This is the LLM-provided guidance.

Base:
https://tts.localhost/v1
Relative path:
"/audio/speech"
Since the second argument starts with a leading slash (/audio/speech), urljoin discards everything after the “network location” (https://tts.localhost) and replaces it with /audio/speech.

That’s why /v1 disappears.

@98h398hrpohpoai commented on GitHub (Sep 9, 2025): This quick fix worked for me (I'm dumb, don't assume this won't break other stuff--I have no idea): Remove the leading slash from the relative path of the join statement outlined below. ``` url=urljoin( request.app.state.config.TTS_OPENAI_API_BASE_URL, "/audio/speech", ``` to ``` url=urljoin( request.app.state.config.TTS_OPENAI_API_BASE_URL, "audio/speech", ``` In line 342 of open_webui/routers/audio.py --- Also ensure the base tts URL ends with a slash (because we can't use a leading slash in the join function) e.g. https://tts.localhost/v1/ This is the LLM-provided guidance. ``` Base: https://tts.localhost/v1 Relative path: "/audio/speech" Since the second argument starts with a leading slash (/audio/speech), urljoin discards everything after the “network location” (https://tts.localhost) and replaces it with /audio/speech. That’s why /v1 disappears. ```

GiteaMirror commented

2026-05-13 05:01:42 -05:00

@tjbck commented on GitHub (Sep 10, 2025):

Reverted with 8339f59cdf

@tjbck commented on GitHub (Sep 10, 2025): Reverted with 8339f59cdfc63f2d58c8e26933d1bf1438479d75

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#72842