OptionalapiAPI key to use. If not provided, it will look for the GOOGLE_GENAI_API_KEY or GEMINI_API_KEY environment variable. If gemini provider is selected and no key is provided, the fake key "-" will be used for the "x-goog-api-key" header.
OptionalheadersHeaders to merge with internally crafted headers.
OptionallocationThe Vertex AI location. Required if vertexai is true.
The name of the model to use. The model string specifies the LLM provider
(e.g., Vertex AI, Gemini), API version, and the model ID. Supported format:
apigee/[<provider>/][<version>/]<model_id>
Components:
provider (optional): vertex_ai or gemini.
version (optional): The API version (e.g., v1, v1beta). If not
provided, a default version will selected based on the provider.
model_id (required): The model identifier (e.g.,
gemini-2.5-flash).
Examples:
- apigee/gemini-2.5-flash
- apigee/v1/gemini-2.5-flash
- apigee/vertex_ai/gemini-2.5-flash
- apigee/gemini/v1/gemini-2.5-flash
- apigee/vertex_ai/v1beta/gemini-2.5-flash
OptionalprojectThe Vertex AI project ID. Required if vertexai is true.
OptionalproxyThe proxy URL for the provider API. If not provided, it will look for the APIGEE_PROXY_URL environment variable.
OptionalvertexaiWhether to use Vertex AI. If true, project, location
should be provided.
The parameters for creating a Gemini instance.