Gemini (LLM)

Use Gemini as the LLM (dialogue + reasoning) provider.

Setup

• Set GEMINI_API_KEY

# .env
GEMINI_API_KEY=...

Example

from siphon.plugins import gemini

llm = gemini.LLM(
    model="gemini-2.0-flash",
)

Common options

• model (default: gemini-2.5-flash-lite)
• api_key (default: None)
• max_output_tokens (default: 150)
• temperature (default: 0.3)