An open-source AI agent that brings the power of Gemini directly into your terminal.
Gemini CLI is an open-source terminal-based AI agent that integrates Google's Gemini models into the command line, enabling developers to perform code understanding, generation, and automation tasks. The project is built with TypeScript and Node.js, uses React with the Ink framework for CLI rendering, Vitest for testing, and esbuild for bundling, organized as a monorepo with separate packages for the CLI interface, core backend logic, an experimental Agent-to-Agent server, an SDK, developer tools, and a VS Code extension companion. The repository includes a comprehensive behavioral evaluation framework that tests the model's decision-making in response to specific prompts, distinguishing between tests that should always pass and those that typically pass but may occasionally vary, with a structured promotion process for advancing tests from incubation to full CI inclusion.