Thanks — I’m responding to your question about whether Speechmatics ASR can switch languages mid-conversation in real time. Short answer: That capability is not confirmed. The Speechmatics STT plugin supports multilingual transcription and recognizes different dialects and accents, and the

language

parameter is set per session (default

en

) but can be changed for a session. It’s not stated that Speechmatics will automatically detect and switch between distinct languages (for example English → Spanish) within the same utterance or mid-conversation. If you need a solution, here are some practical options we can explore together: • If you control the session: explicitly set or change the session

language

when speakers switch languages. This requires detecting the switch in your app and restarting or reconfiguring the transcription session with the new

language

. • If speakers frequently code-switch within utterances: Speechmatics may still transcribe accents/dialects, but automatic language detection for real-time code-switching is not confirmed. You may want to test with your target language mix to gauge transcription quality. • Use interim/partial transcripts and speaker diarization (supported) to help identify segments where you might switch languages, then reprocess those segments with the appropriate language if needed. Would you like help with one of these approaches? If so, tell me: • Whether you control the transcription session programmatically (yes/no). • Which languages you expect to see and whether switches happen between utterances or within a single utterance. I can then suggest specific implementation steps or test approaches. Reference: Speechmatics STT plugin guide — https://docs.livekit.io/agents/models/stt/speechmatics Sources: Speechmatics STT plugin guide | LiveKit Documentation, with interactive elements