Claude Code

Claude Code is the command-line agent from Anthropic, launched in beta in February 2026. Unlike classic AI IDEs (Cursor, Windsurf) that stay inside the editor, Claude Code lives in your terminal and can execute real shell commands: run tests, create Git branches, deploy to Vercel, tail production logs.

Usage is trivial: type claude in any project folder, describe what you want ("fix the checkout bug and add a test"), and Claude runs every step. It asks for confirmation before destructive actions (rm, git push).

Claude Code is included with Claude Pro ($20/mo) and Max ($100/mo) plans. It can also run on an Anthropic API key in pay-per-use mode.

Key features

• Terminal-native: lives in your shell, runs commands.
• Agentic autonomy: chains reading, writing, testing, deploying.
• Permissions: explicit confirmation on destructive actions.
• Skills & Hooks: reusable workflows and triggers.
• Sub-agents: parallel tasks for independent explorations.
• MCP: connect to Model Context Protocol servers.