Source: Gemini Live API Docs

Google's overview of the Gemini Live API: low-latency streaming voice and vision over a stateful WebSocket, audio formats, barge-in, tools and ephemeral tokens.

Contents
  • What: official API documentation (Live API overview)
  • Author/publisher: Google
  • Date: living document; read 2026-11-14
  • Original: Live API

Summary

The Live API streams audio, images and text into Gemini and returns spoken responses in real time, over one stateful WebSocket.

Key claims

  • Input audio: raw 16-bit PCM, 16 kHz, little-endian. Output audio: raw 16-bit PCM, 24 kHz. Image input: JPEG up to 1 frame per second.
  • Protocol: stateful WebSocket (WSS). Client-to-server is simpler and usually streams better; use ephemeral tokens rather than API keys in production. Server-to-server proxies through your backend.
  • Features: barge-in, function calling and Google Search, transcripts of both sides, proactive audio, affective dialog, 70 languages, live translation.
  • Partner integrations include LiveKit and Pipecat.

Feeds

Voice agents · Multimodal models

Suggest changes →