Skip to main content

Google Provider

💡TL;DR

The Google preset uses the native Gemini Generative Language API. Configure an API key authorized for that endpoint and choose a currently available model. A gateway exposing Gemini through OpenAI-compatible chat uses the gateway preset instead.

Setup​

Select Google, enter the key, retain the native Generative Language endpoint, and choose a model served by that API version. Run Test connection, then translate or process a disposable note. The preset does not configure an arbitrary Google Cloud or Vertex deployment automatically.

Endpoint And Authentication​

FieldPreset value
Transportgoogle
API keyRequired
Base URLhttps://generativelanguage.googleapis.com/v1
Default modelgemini-2.0-flash-exp
Temperature0.5
Connection testchat-only

The runtime constructs a native model generateContent request with API-key authentication. The preset's historical model ID is not a current-availability guarantee. Review endpoint URLs before sharing diagnostics because authentication information can appear in request details.

Model Discovery​

Use the native Google model discovery when available. Confirm that the chosen model supports generation on the configured API version. An empty list can indicate permissions, endpoint or model-availability problems; test the actual configuration rather than changing protocols at random.

Troubleshooting​

  • 400: inspect model ID, request parameters and token/output limits.
  • 401 or 403: check API-key restrictions, API enablement and model entitlement.
  • 404: check whether that model exists on the selected API version.
  • Gateway examples fail: use the corresponding OpenAI-compatible gateway preset.

When To Use​

Use this preset for the native Gemini API. Evaluate the intended translation, research or diagram task on a representative sample. This guide does not promise a model's relative speed, cost or translation quality.

Next Steps​