Google Gemini

Sinaptic® DROID+ supports Google's Gemini models via the Generative Language API.

Setup

1. Get an API key from Google AI Studio
2. Add it to your environment or .env file:

export GEMINI_API_KEY=AIza...

3. Configure in droid.yaml:

gemini:
  base_url: "https://generativelanguage.googleapis.com/v1beta"
  api_key: "${GEMINI_API_KEY}"

Agent Config

name: "gemini-agent"
model:
  provider: "gemini"
  name: "gemini-2.0-flash"
  max_tokens: 2048
  temperature: 0.7

Available Models

Model	Best for	Context
gemini-2.0-flash	Fast, multimodal (recommended)	1M
gemini-2.0-flash-lite	Fastest, cost-effective	1M
gemini-1.5-pro	Complex reasoning	2M

Check Google's model list for the latest.

Notes

• Gemini models support function calling, so all Sinaptic® DROID+ tools work with them.
• The Generative Language API is free-tier friendly — Google offers generous free quotas for Gemini Flash models.
• Sinaptic® DROID+ translates between the OpenAI request format and Gemini's API internally. Your clients use the standard OpenAI SDK.