Claude Code
Terminal AI agent for complex codebases
What it is
Anthropic's official CLI that turns Claude into an agent that can read, write, execute commands, and commit to your local repository. Unlike editor plugins, it operates on the entire codebase without context window limits. Ideal for large refactors, complex debugging, and features spanning dozens of files.
How to use it, step by step
- 1
Install Node.js, then run from a terminal: npm install -g @anthropic-ai/claude-code.
- 2
Move into your project folder (cd projectname) and run: claude.
- 3
On first launch, log in with your Anthropic account (or enter an API key). Claude reads the project structure on its own.
- 4
Describe what you want in plain language: 'add tests to this module', 'find and fix the login bug'. The agent reads files, proposes changes, and asks for confirmation before writing or running commands.
- 5
Approve actions one by one; when done you can ask it to commit with a sensible message.
💡Practical tips
- →Create a CLAUDE.md file in the root with permanent instructions (style, test commands): the agent reads it on every launch.
- →For big jobs ask for a plan first ('make a plan, don't write code yet'), then have it execute step by step.
- →Works best inside a Git repo: that way you can always roll back changes with git.
💰Pricing
Pay-as-you-go via the Anthropic API, or included in Claude Pro/Max plans (from about $20/month) with usage limits.
The official site opens in a new tab.