Intelligent transcription with Gemini 3.5 Transcribe
Source
blog.google
Date
Context
Google AI's Gemini 3.5 Transcribe offers different interfaces for live speech and recorded audio. The announcement describes continuous transcription through the Live API, while the Interactions API processes recordings with speaker attribution and word-level timestamps. Those recorded-audio features help connect a passage of text to who spoke and where it occurred.
The model also supports vocabulary supplied by the developer and cleanup of fillers or spoken corrections. Developers choosing an interface therefore need to distinguish immediate dictation from a transcript organized for later review. The announcement specifically places speaker attribution and word-level timing in the recorded-audio workflow.
