Skip to Content
Model API ReferenceGeminiOverview

Gemini API

Use https://api.fenno.ai as the host for native Gemini requests. The Google GenAI SDK adds its version and model-resource paths.

MethodFull pathReference
POST/v1beta/models/{model}:generateContentGenerate content
POST/v1beta/models/{model}:streamGenerateContent?alt=sseStreaming content
GET/v1beta/modelsModel resources
GET/v1beta/models/{model}Model resources

Authentication

x-goog-api-key: <FENNO_API_KEY>
Content-Type: application/json

Prefer the header to a key query parameter to reduce credential exposure in logs. SDKs may manage authentication transport themselves; configure them without printing request URLs containing credentials.

Protocol boundaries

  • Input uses contents[].parts, not OpenAI messages or Anthropic content blocks.
  • REST generation settings use generationConfig; SDK helpers can expose these through config.
  • Output is in candidates[].content.parts; inspect finish reasons and safety feedback too.
  • Do not infer support for Files, cachedContents management, tuning or every native hosted tool from content-generation support.

See Gemini CLI setup, vision inputs, and the official Gemini API reference .

Token countingGenerate Content and streaming