[Gemini 3.5 Transcribe](https://vercel.com/ai-gateway/models/gemini-3.5-transcribe-live) from Google is now available on AI Gateway. It takes audio and returns text, in two variants:
`google/gemini-3.5-transcribe`
transcribes a complete recording in a single request.
google/gemini-3.5-transcribe-live
transcribes audio over a WebSocket, returning a transcript that updates while the recording is still going.
The model detects the language on its own, covers 85+, and follows a speaker who switches language partway through. You can also supply custom vocabulary so it recognizes names, jargon, and spellings.
Streaming transcription is new in AI SDK V7:
streamTranscribe
opens the socket and takes a ReadableStream
of raw audio chunks, so you can pass a microphone straight through. Tell it the format you are sending with inputAudioFormat
:
For audio you already have on disk, transcribe
sends it in one request and returns the text:
You can also try the model without writing any code. Open Gemini 3.5 Transcribe Live and send audio to read the transcript in the browser.
AI Gateway provides a unified API for calling models, tracking usage and cost, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, budgets for API keys, routing rules, and more.
AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests.
You can view all transcription models available on AI Gateway, or start from the speech quickstart.