What is Claude Code?

Claude Code is Anthropic's AI coding agent — a tool that lives in your terminal, reads your whole codebase, and turns plain-English requests into real code changes.

The workflow feels different from autocomplete-style assistants. You type something like "add rate limiting to the API and write tests for it," and Claude Code figures out which files matter, makes the edits, runs the test suite, fixes what breaks, and can commit the result — showing its work and asking permission before anything risky.

Launched in early 2025, it did a lot to prove that a coding agent in a terminal — no fancy editor UI at all — could be a serious daily tool. A few things people like about it:

• It works where you already work. Terminal first, with extensions for editors like VS Code and JetBrains.
• It's deeply customizable. A CLAUDE.md file teaches it your project's conventions, and skills, hooks, and MCP servers extend what it can do.
• It handles long jobs. Subagents and background tasks let it grind through multi-step work without constant babysitting.

The trade-off: it's tied to Anthropic's Claude models, and heavy use costs real money — either usage-based API billing or one of Claude's paid plans.

Its closest rivals are OpenAI's Codex and Cursor's agent mode. See Cursor vs Copilot vs Claude Code for how they stack up.