Expand description
§inference-runtime-gemini
Google Gemini runtime for both AI Studio (api_key) and Vertex AI
(OAuth2 over project + region). Doc §10.3.
OAuth2 token refresh is exposed as a pluggable
CredentialProvider
trait so we don’t pull a full oauth2 stack into the workspace
root. Operators wire StaticApiKey for AI Studio or supply their
own provider for Vertex (typically gcloud auth print-access-token
refreshed on a timer).
Re-exports§
pub use config::GeminiConfig;pub use config::GeminiVariant;pub use config::SafetySetting;pub use cost::GeminiPricing;pub use error::classify_gemini_error;pub use runner::GeminiRunner;