LiveKit Community

When using Speechmatics
```end_of_utterance_silence_trigger=0.2,
max_delay=0.8,
end_of_utterance_mode=EndOfUtteranceMode.FIXED```
, should I also use Silero VAD?

Thanks for reaching out, Eray. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.


Thanks — regarding your Speechmatics settings (`end_of_utterance_silence_trigger=0.2`, `max_delay=0.8`, `end_of_utterance_mode=EndOfUtteranceMode.FIXED`):

Yes — you should still use Silero VAD. Speechmatics' EOU settings control final transcript-based endpointing, but LiveKit relies on a VAD plugin like Silero for low-latency voice/silence detection and interruption handling. Using Silero VAD together with Speechmatics improves responsiveness (detecting mid-utterance speech or quick restarts) while Speechmatics provides accurate final EOU decisions.

If you want, tell me your main goal (lower latency, better interruption handling, or fewer false endpoints) and I’ll suggest exact VAD + Speechmatics tuning.
 Sources: <https://docs.livekit.io/agents/logic-structure/turns/vad/|Silero VAD plugin | LiveKit docs> | <https://docs.livekit.io/agents/models/stt/plugins/speechmatics/|Speechmatics STT plugin guide | LiveKit docs> | <https://docs.livekit.io/agents/logic-structure/turns/|Turns overview | LiveKit docs>