gemini-cli

Gemini CLI is an open-source AI agent from Google that brings the Gemini model into your terminal — the text-based window where developers type commands. Instead of switching to a browser to chat with an AI, you run a single command and start a conversation that already has access to your files, your shell, and tools for searching the web. The README describes it as the most direct path from your prompt to Google's Gemini model. It installs through npm, Homebrew, MacPorts, or Anaconda, or you can launch it instantly with npx. Inside, it acts as an agent: it can read and edit files in a codebase, run shell commands, fetch pages from the web, and use Google Search to ground its answers in real-time information. It supports the Model Context Protocol (MCP), an open standard for plugging extra tools and data sources into an AI assistant. Conversation checkpoints save and resume long sessions, and project-specific GEMINI.md files tailor behavior per repository. A non-interactive mode takes a prompt and returns plain text or structured JSON for shell scripts, plus a companion GitHub Action runs pull-request reviews and issue triage in CI. You would use Gemini CLI if you spend most of your day in the terminal and want an AI pair-programmer that can actually touch your code, if you want to script Gemini into a build pipeline, or if you want to wire it into GitHub workflows. It offers three authentication paths — Google sign-in with a free tier, a Gemini API key, or Vertex AI for enterprise use. The project is written in TypeScript and licensed under Apache 2.0.