Official documentation
Live transcription with Gemini Live API
Google, Google AI for Developers, 2026
Read the originalRead on October 17, 2026
- What it is
- Google's guide to streaming audio over WebSockets to the gemini-3.5-transcribe-live model, which returns incremental text as speech occurs. It lists the session limit and the features the streaming model does not support.
- How much weight it carries
- Official Google documentation, read on 2026-10-17. Session limits and model names can change. The page names a model that we did not find on the Vertex AI Live API overview, so check which platform offers it before you build on it.
Facts taken from this source
- A Gemini live transcription session supports continuous streaming for up to 10 minutes.
“Live transcription sessions support continuous streaming for up to 10 minutes.” (Limitations)
- Speaker diarization is not supported in Gemini live transcription sessions, and Google points to its non-streaming audio transcription endpoint for speaker labels.
“Speaker diarization is not supported in live streaming sessions. For speaker diarization, use the non-streaming Audio transcription endpoint.” (Limitations)