Most of us have used Claude in a chat window. But Anthropic also ships
Claude Code — an AI coding agent that works from your terminal, reads
your actual project, and edits real files. It's a different way of working,
and I wanted to write down what I learned.
The core difference
When you ask an AI in a chat window, it only sees the code you paste in.
Claude Code reads the whole project structure through the terminal, so it
can tell you which files need to change before you touch anything —
database, controllers, routes, UI, validation, permissions.
What this article covers
- What Claude Code can do — explain code, debug across the stack, refactor, add features, and support your Git workflow.
- A real example — I ran it against a multi-tenant quotation platform had it analyze the whole structure Real output screenshots included.
-
Installation and slash commands —
/init,/plan,/diff,/context, and advanced ones like/goal. - Who it's for, plus a recommended workflow: analyze → plan → edit → review diff → test → commit.
- What to watch out for — commit before letting AI edit, review every diff, never paste API keys.
One thing worth knowing upfront: Claude Code needs a paid plan (Pro, Max,
Team, Enterprise, or Console API). The free Claude plan doesn't include it.
Full article (available in English and Thai):
Top comments (0)