mirror of
https://github.com/open-webui/open-webui.git
synced 2026-07-23 03:23:08 -05:00
issue: speech to text broken in 0.6.6 #5047
No Assignees
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: github-starred/open-webui#5047
Reference in New Issue
Block a user
Originally created by @morbificagent on GitHub (May 5, 2025).
Check Existing Issues
Installation Method
Docker
Open WebUI Version
0.6.6
Ollama Version (if applicable)
No response
Operating System
Ubuntu
Browser (if applicable)
Kiwi mobile
Confirmation
README.md.Expected Behavior
Speech to text should work
Actual Behavior
When I press the microphone button and record audio I get the error message
"Oops! It seems like the file format you're trying to upload is not supported. Please upload a file with a supported format and try again."
after ending the recording.
In 0.6.5 it was working without problem
Steps to Reproduce
Logs & Screenshots
Oops! It seems like the file format you're trying to upload is not supported. Please upload a file with a supported format and try again.
Additional Information
No response
@NoMoreFood commented on GitHub (May 5, 2025):
I doubt this is my change unless you're using Azure AI STT, but could you post the output from your console session?
@maxkuminov commented on GitHub (May 5, 2025):
I have the same issue on both my desktop (linux/firefox) and Android phone (Chrome PWA).
The server logs show:
2025-05-05 16:03:31.411 | INFO | open_webui.routers.audio:transcription:790 - file.content_type: audio/webm;codecs=opus - {}
2025-05-05 16:03:31.414 | INFO | uvicorn.protocols.http.httptools_impl:send:476 - 192.168.2.38:0 - "POST /api/v1/audio/transcriptions HTTP/1.1" 400 - {}
It looks like the front end is trying to record audio in the webm format, and then send it to OpenAI, and this is an unsupported format.
@k1rb commented on GitHub (May 5, 2025):
Just updated to 0.6.6 having the same issue. Fedora, firefox, using the mic input on open-webui, immediate error.
"Oops! It seems like the file format you're trying to upload is not supported. Please upload a file with a supported format and try again."
I have same console results as above.
@dguembel-itomig commented on GitHub (May 5, 2025):
Same problem here. Chromium as the Browser. Also happens when using "local" Whisper.
@mreitschuster commented on GitHub (May 5, 2025):
same here in chats. but in "notes" the recording works (and correctly extracts the text from speech). also call functionality correctly does speech to text. Just the "record voice" doesnt. i am using "Whisper (local) " as SST model
@tjbck commented on GitHub (May 5, 2025):
Fixed with
7b36466c1c@maglat commented on GitHub (May 5, 2025):
The fix helped to accept the audio, but after finishing the recoding, openwebui crashs.
`2025-05-05 23:33:44 2025-05-05 21:33:44.477 | INFO | open_webui.routers.audio:transcription:790 - file.content_type: audio/webm; codecs=opus - {}
2025-05-05 23:33:44 2025-05-05 21:33:44.478 | INFO | open_webui.routers.audio:transcribe:499 - transcribe: /app/backend/data/cache/audio/transcriptions/bdf91f1a-f1bb-4511-8206-1c384443f6a7.webm - {}`
No error in the log. just simply stops the docker container
@morbificagent commented on GitHub (May 6, 2025):
It's working with dev and docker here
@Xi-Gong commented on GitHub (May 6, 2025):
Same issues here, patch 7b36466 doesn't fix the problem for open-webui install from pip.
@ITmail-dk commented on GitHub (May 7, 2025):
Temporary fix, edit file in the container:
See CONTAINER-ID via docker ps
docker exec -i -t CONTAINER-ID bash
sed -i 's/supported_filetypes = ("audio/mpeg", "audio/wav", "audio/ogg", "audio/x-m4a")/supported_filetypes = ("audio/mpeg", "audio/wav", "audio/ogg", "audio/x-m4a", "audio/webm")/' open_webui/routers/audio.py
@lnastjsj commented on GitHub (May 12, 2025):
0.6.9 Engine OpenAl "record voice"Report an error:[ERROR: 400: [ERROR: Open WebUI: Server Connection Error]]
@AlexUs5 commented on GitHub (Jun 18, 2025):
Again same issue in 0.6.15
"Oops! It seems like the file format you're trying to upload is not supported. Please upload a file with a supported format and try again."
@NoMoreFood commented on GitHub (Jun 18, 2025):
@AlexUs5 Try adding in audio/,video/ as supported content types?
@AlexUs5 commented on GitHub (Jun 18, 2025):
@NoMoreFood "Settings- Audio - Supported MIME Types" empty by default. Now i added "audio/mpeg, audio/wav, audio/ogg, audio/x-m4a, audio/webm" but again have error "Oops! It seems like the file format you're trying to upload is not supported. Please upload a file with a supported format and try again."
@jwhb commented on GitHub (Jun 18, 2025):
Thanks. To be precise, this should work:
Supported MIME Types:
audio/*,video/*@AlexUs5 commented on GitHub (Jun 19, 2025):
@jwhb
Thanks. Now all works
@rikrdo89 commented on GitHub (Jul 10, 2025):
@jwhb that is the solution!
@dollro commented on GitHub (Aug 2, 2025):
@jwhb : works! perfect!
Can we ensure that this is set per default? so many struggle here to get this right!