Skip to main content

Usage guide

Setup, environment variables, and Python/JavaScript/Go usage examples.

​
GeminiRealtime

Speech-to-speech conversation with the Gemini Live API, e.g. “gemini-3.8-live”. Use it when one model should listen, reason and talk with minimal latency. Authenticates with GOOGLE_API_KEY from your environment, or an explicit api_key.

​
Fields

​
str | None
default:"UNSET"
Gemini Live model name, e.g. “gemini-3.8-live”. Defaults to ‘gemini-3.8-live’.
​
Mapping[str, Any] | None
default:"UNSET"
Model settings as a dict, e.g. voice (“Puck”, “Kore”), language_code, temperature, max_output_tokens, response_modalities and transcription options. Defaults to None (not set).
​
str | None
default:"UNSET"
Path to a Google Cloud service account JSON file, used instead of an API key. Defaults to None (not set).
​
bool
default:"UNSET"
Connect through Vertex AI on Google Cloud instead of the Gemini API. Defaults to False.
​
Mapping[str, Any] | None
default:"UNSET"
Vertex AI connection settings as a dict, such as the Google Cloud project and location. Used when vertexai is on. Defaults to None (not set).