Gemini Code is a terminal-based AI coding assistant powered by Google's Gemini 2.5 Pro, designed for developers seeking an integrated, interactive coding partner. It automates common development tasks through intelligent tool usage, enhancing productivity within the terminal environment.

How It Works

The assistant leverages Gemini's native function calling capabilities to automatically select and execute relevant tools for tasks like file operations, directory management, system commands, linting, and test execution. This approach mimics natural language interaction, allowing users to request actions without explicitly specifying commands, leading to a more intuitive workflow.

Quick Start & Requirements

• Install via pip: pip install gemini-code
• Requires a Google API key for Gemini models. Setup: gemini setup YOUR_GOOGLE_API_KEY
• Official documentation: https://github.com/raizamartin/gemini-code

Highlighted Details

• Supports multiple Gemini models, including Gemini 2.5 Pro and 1.5 Pro.
• Features automatic tool usage for file/directory operations, system commands, linting, and test running (e.g., pytest).
• Implements native Gemini function calling for reliable tool integration.
• Includes interactive commands like /exit and /help.

Maintenance & Community

The project is under active development, with recent updates focusing on enhancing tool execution, error handling, and model integration. Community links are not specified in the README.

Licensing & Compatibility

• License: MIT
• Compatible with commercial use and closed-source linking.

Limitations & Caveats

Users who previously created a configuration file may need to delete it (rm -rf ~/.config/gemini-code) to ensure correct defaults with newer versions.