Claude Code

Anthropic's command-line tool for coding with Claude. Runs in your terminal, edits files in your project, and remembers context across sessions.

What it is

Claude Code is Anthropic's CLI for coding alongside their Claude models. You install it once (npm i -g @anthropic-ai/claude-code), point it at any project folder, and ask it to read, write, refactor, or ship code. It's not a chatbox in a sidebar. It's a terminal agent that can run shell commands, read your files, edit them in place, and persist memory between runs.

Why CEOs care

This is the harness most operator-grade workflows now sit on. If your team builds, your engineers are likely deciding between Claude Code, Cursor, and a few others right now. Understanding the difference matters because Claude Code is what unlocks "leave it running overnight" autonomous routines · the kind of work that compounds while the desk is empty.

When you'll encounter it

In any technical hire's setup, in workflow articles on this site, in conversations about agents and autonomous routines, and increasingly in non-engineering teams that have figured out how to point it at their docs and ops scripts.

Example

Andrew runs Claude Code as the harness behind Headphones.com's daily ops routines · pulling sales data, tagging support tickets, drafting weekly reports. The same install scaffolds new product specs and writes migrations. One tool, ~12 hours/week of leverage.

Related

• harness
• mcp-server
• autonomous-routine