Claude Code

Install it with one command (curl -fsSL https://claude.ai/install.sh | bash on macOS, Linux and WSL; Homebrew, WinGet and a PowerShell script are also documented), then run claude inside a repository and describe the task in plain language. It searches the codebase itself instead of asking you to pick context files, proposes a plan, makes the edits and runs commands, pausing for approval on anything your permission settings do not allow.

• •CLAUDE.md project memory - A markdown file in your project root that Claude Code reads at the start of every session, so build commands and conventions carry over.
• •MCP servers - Connects to external tools and data sources through the Model Context Protocol, the open standard for plugging AI tools into other systems.
• •Hooks and skills - Hooks run shell commands before or after its actions, like formatting after every edit; skills package repeatable team workflows such as /review-pr.
• •Subagents - Spawns several Claude Code agents that work on different parts of a task at the same time.
• •Headless and CI - Scriptable with claude -p for pipes and automation, and runs in GitHub Actions or GitLab CI/CD for code review and issue triage.

Developers who want to hand off whole tasks rather than accept autocomplete: getting oriented in an unfamiliar repository, working through an issue queue, or starting the migration nobody has had time for. Terminal-first engineers get the most out of it. If you mostly want inline suggestions while you type, Cursor or GitHub Copilot fit better.