Chenghao Mou
Chenghao Mou
We have a PR #4098 that should fix this (similar issue #4048)
There is another similar issue: when dealing with false interruption, `agent_speaking` (after the pause) is grouped under a `user_turn`. I am open to solving it in a different PR.
Yes, VAD interruptions work like you described. > maybe the feature would be interruption based on how many words the agent said : like if the agent said less than...
> But while user is speaking, the agent transitions from thinking -> speaking but immediately after 100ms after uttering a blip of audio "I" it goes back into listening. It...
> > Something is off here. Whenever I take a pause longer than a few seconds, the connection will throw a `APIStatusError(message="ElevenLabs STT connection closed unexpectedly")` with a `WSMessage(type=, data=1000,...
Yeah, I agree. Should we just add a warning somewhere in the example or readme? I think it is totally fine to have the implementation available.
> @chenghao-mou update: ~it seems it elevenlabs STT works when server VAD is disabled~ it's better when server VAD is disabled, but still sometimes got some random output from STT,...
Here is one way to pinpoint the issue: 1. In your local `/home/sarangmohod/livekit/livekit-voice-agent/.venv/lib/python3.10/site-packages/livekit/agents/worker.py`, locate the function call around line 695. Add print statements before that to see the actual api...
I can see in the code that `participant_metadata_changed` is supposed to be an internal event, not user-facing. I do see metadata being reversed with your code. I would suggest using...