github/copilot-cli

GitHub Copilot CLI is an AI-powered coding assistant that runs directly in your terminal. Instead of opening a browser or switching to a separate app, you type natural language requests in your command line and an AI agent helps you build, debug, understand, and refactor code. It works on Linux, macOS, and Windows. The tool requires an active GitHub Copilot subscription, which is a paid service from GitHub. Once installed and authenticated with your GitHub account, you launch it by typing copilot in any folder containing code you want to work with. From there you can have a conversation with the AI about your code, ask it to make changes, or get explanations of what existing code does. A key feature is that nothing runs automatically without your approval. The agent can plan and propose multi-step changes, but you review each action before it executes. It also comes connected to GitHub's own infrastructure by default, so it can access your repositories, issues, and pull requests using plain language. The CLI supports extending its capabilities through a standard called MCP (Model Context Protocol), which lets you plug in additional tools or data sources beyond what ships in the box. It also has optional support for Language Server Protocol, which is a way to give the AI more precise understanding of code in specific programming languages by connecting it to a language analysis tool you install separately. Installation is straightforward on all platforms: a one-line curl command on Mac and Linux, Homebrew, WinGet on Windows, or npm. The default AI model is Claude Sonnet, but you can switch models using a command inside the tool. The project is actively developed, with frequent updates, and the GitHub team is collecting feedback through the repository's issue tracker.