No documentation on voice mode #112

Closed
opened 2026-02-15 17:15:06 -05:00 by yindo · 6 comments
Owner

Originally created by @lukaas33 on GitHub (Nov 13, 2025).

Next to the microphone button there is a button that starts what I think is called voice mode.
I could not find any documentation on this feature but I assume it does realtime interaction with models that support it.
For me this feature does not work as expected since it does not respond to any of my microphone input.
For reference, both TTS and STT work for me.
My guess of why this doesn't work is because litellm does not seem to support the realtime endpoints.
I am fine with it not working but would like to know how to remove the button from the GUI then.

More information in the docs on this would be appreciated.

Originally created by @lukaas33 on GitHub (Nov 13, 2025). Next to the microphone button there is a button that starts what I think is called voice mode. I could not find any documentation on this feature but I assume it does realtime interaction with models that support it. For me this feature does not work as expected since it does not respond to any of my microphone input. For reference, both TTS and STT work for me. My guess of why this doesn't work is because litellm does not seem to support the realtime endpoints. I am fine with it not working but would like to know how to remove the button from the GUI then. More information in the docs on this would be appreciated.
yindo closed this issue 2026-02-15 17:15:06 -05:00
Author
Owner

@Classic298 commented on GitHub (Nov 13, 2025):

yes you need a realtime compatible TTS engine like TTS-1 or TTS-HD by openAI

it needs streaming/realtime.

@Classic298 commented on GitHub (Nov 13, 2025): yes you need a realtime compatible TTS engine like TTS-1 or TTS-HD by openAI it needs streaming/realtime.
Author
Owner

@Classic298 commented on GitHub (Nov 13, 2025):

Not sure where/how to document this as there is no extra page for this feature hmm..

Maybe I will add one or add more documentation to related environment variables once this gets merged: https://github.com/open-webui/open-webui/pull/18607

@Classic298 commented on GitHub (Nov 13, 2025): Not sure where/how to document this as there is no extra page for this feature hmm.. Maybe I will add one or add more documentation to related environment variables once this gets merged: https://github.com/open-webui/open-webui/pull/18607
Author
Owner

@Classic298 commented on GitHub (Nov 14, 2025):

just got implemented lol. Will pin this and work on it soon

@Classic298 commented on GitHub (Nov 14, 2025): just got implemented lol. Will pin this and work on it soon
Author
Owner

@Classic298 commented on GitHub (Nov 14, 2025):

@silentoplayz you have a proposal for me? Should we add this to the features section?

@Classic298 commented on GitHub (Nov 14, 2025): @silentoplayz you have a proposal for me? Should we add this to the features section?
Author
Owner

@silentoplayz commented on GitHub (Nov 14, 2025):

I believe it should be documented, yes. I think it could be its own page in the STT/TTS section.

Image
@silentoplayz commented on GitHub (Nov 14, 2025): I believe it should be documented, yes. I think it could be its own page in the STT/TTS section. <img width="284" height="156" alt="Image" src="https://github.com/user-attachments/assets/7676bb2d-e87b-4a60-bb0f-1af0d1a20638" />
Author
Owner

@Classic298 commented on GitHub (Nov 23, 2025):

@lukaas33

I am fine with it not working but would like to know how to remove the button from the GUI then.

There is a permission toggle for this in the admin panel

I could not find any documentation on this feature but I assume it does realtime interaction with models that support it.
For me this feature does not work as expected since it does not respond to any of my microphone input.
For reference, both TTS and STT work for me.
My guess of why this doesn't work is because litellm does not seem to support the realtime endpoints.

its already documented and the env vars for this feature were recently expanded.
and the voice mode is used by setting an LLM, TTS and STT model.

Realtime models are not supported yet.
Feature request for realtime models is here: https://github.com/open-webui/open-webui/issues/5894

@Classic298 commented on GitHub (Nov 23, 2025): @lukaas33 > I am fine with it not working but would like to know how to remove the button from the GUI then. There is a permission toggle for this in the admin panel > I could not find any documentation on this feature but I assume it does realtime interaction with models that support it. > For me this feature does not work as expected since it does not respond to any of my microphone input. > For reference, both TTS and STT work for me. > My guess of why this doesn't work is because litellm does not seem to support the realtime endpoints. its already documented and the env vars for this feature were recently expanded. and the voice mode is used by setting an LLM, TTS and STT model. Realtime models are not supported yet. Feature request for realtime models is here: https://github.com/open-webui/open-webui/issues/5894
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: open-webui/docs#112