docs/api/realtime/google.md
pydantic_ai.realtime.googleThe Gemini Live API provider. Requires the google optional group
(pip install "pydantic-ai-slim[google]").
[GoogleRealtimeModel][pydantic_ai.realtime.google.GoogleRealtimeModel] runs over the google-genai
SDK (which manages the WebSocket transport). Gemini expects 16 kHz PCM input (output is 24 kHz),
produces one response modality per session, and natively accepts live video frames sent as
[BinaryImage][pydantic_ai.messages.BinaryImage]. It exposes Gemini Live's session and generation
configuration through [GoogleRealtimeModelSettings][pydantic_ai.realtime.google.GoogleRealtimeModelSettings] —
shared turn-taking via [TurnDetection][pydantic_ai.realtime.TurnDetection], with finer Gemini-specific
control via [AutomaticVAD][pydantic_ai.realtime.google.AutomaticVAD] in google_vad plus
google_activity_handling/google_turn_coverage, voice via google_voice or a
[MultiSpeaker][pydantic_ai.realtime.google.MultiSpeaker] in google_multi_speaker,
and long-session [ContextCompression][pydantic_ai.realtime.google.ContextCompression] — with
resilience via session resumption + a [ReconnectPolicy][pydantic_ai.realtime.ReconnectPolicy] in the
reconnect setting.
::: pydantic_ai.realtime.google