What Claude Code is

Most AI coding help you may have seen is a chat window: you paste code in, copy an answer out. Claude Code works differently. It is an agent that sits inside your project. You describe a goal in plain language, such as "the signup form accepts empty emails, fix it", and it looks around your files, changes code, runs your tests and reports back.

It is made by Anthropic and uses Anthropic's Claude models. The same engine is available in several places, which the docs call surfaces:

• Terminal: the claude command-line program, the most complete version.
• IDE extensions: a panel inside VS Code (and VS Code forks such as Cursor) and a plugin for JetBrains IDEs.
• Desktop app: a standalone app with a Code tab, useful for visual diff review and running several sessions side by side.
• Web: claude.ai/code, where sessions run on cloud machines instead of your computer.

Your project's settings and instruction files work on every surface, so what you learn in the terminal carries over.

What it can do: explain an unfamiliar codebase, fix bugs, write tests, refactor across many files, resolve merge conflicts, write commit messages and open pull requests. What it cannot do: guarantee the result is correct. It is a capable colleague who can still misunderstand you or make mistakes, so you stay responsible for reviewing what it changes.

You need an account to use it: a paid Claude subscription, a Claude Console (API) account, or access through a supported cloud provider. This curriculum focuses on the terminal version, then covers the IDE, CI and SDK.