Gemini CLI

Google's Gemini CLI provides access to the Gemini family of models, which support large context windows (1M+ tokens) and multimodal input.

Setup

1. Install the CLI: Follow the official Google Gemini CLI documentation.

2. Authenticate: Run this in your terminal:

   gemini auth login

3. Verify it works: Test in your terminal:

   gemini "hello"

4. Configure in Burp: Open Burp AI Agent → Settings → AI Backend and set:

Setting	Value
Preferred Backend	Gemini CLI (select from dropdown)
Gemini CLI Command	gemini (or the full path, e.g., /usr/local/bin/gemini)

To specify a model, include the flag in the command field:

Setting	Value
Gemini CLI Command	gemini --model gemini-2.0-flash

Troubleshooting

• "command not found": Ensure gemini is installed and in your system PATH. Try using the full path to the binary.

• Authentication errors: Run gemini auth login in your terminal and follow the OAuth flow.

• Rate limiting: The free tier has rate limits. Consider upgrading to a paid tier for heavy scanning use.