What is Claude Code?

You have a codebase open in your editor and a question about a feature you did not write. In Claude.ai, you would copy the file, paste it into the chat, read the answer, then hop back to your terminal to apply whatever came out of it. Claude Code collapses that loop. It lives next to your files and your shell — and it does the work itself.

Claude Code is not Claude.ai

Claude Code is an agentic coding tool that reads your codebase, edits files, runs terminal commands, and integrates with your existing developer tools. It runs in your terminal, in VS Code, in JetBrains IDEs, in the Claude Desktop app, and on the web. For this course you will use it in the terminal.

The key difference from Claude.ai: direct access. Claude.ai is a chat window — text in, text out. Claude Code has your files, your terminal, and your entire project. No more copy-paste round trips.

What "agentic" actually means

The easiest way to describe the difference: Claude Code works as an AI agent.

An AI agent is software that interacts with its environment and takes actions to complete a defined goal. Under the hood, this is a large language model operating in a loop — call the model, take an action, observe the result, call the model again. Agents reach their goal by using tools, external services, and sometimes other agents.

That loop is what separates "ask the model a question" from "ask the model to ship a feature."

The four things Claude Code can do

Concretely, Claude Code can:

1. Read and understand your codebase. Ask it to explain a feature or trace a bug end-to-end across files.
2. Edit files across your project. Refactor a function and update every call site in a single pass.
3. Run terminal commands. Execute build scripts, run tests, install packages, and use the output to decide what to do next.
4. Search the web. Pull current API docs, release notes, or reference material without leaving the session.

Any one of these is useful. The agent loop combining all four is where the productivity comes from.

Three concepts for using it effectively

Before your first session, load these three ideas into your own working memory. They explain most of what makes Claude Code feel magical — and most of what makes it feel frustrating when you skip them.

Context window — Claude's working memory. It can hold a lot, but not everything. The whole point of "agentic" behaviour is that Claude finds answers inside your codebase without having to load the entire codebase at once. You will manage this deliberately in later lessons.

Permission-based execution. By default, Claude Code asks before running commands or editing files. You are always in control. You can stay hands-on and approve every step, or loosen the reins as you learn to trust it.

It makes mistakes. Claude Code is not perfect. It will sometimes misunderstand your intent, introduce a bug, or over-engineer a solution. Staying in the loop is how you catch those early.

Recap

Claude Code is an agentic coding tool. It reads your codebase, edits your files, runs commands, and connects to external tools so you can ship faster. Unlike Claude.ai, it operates directly on your project instead of trading text through a chat window. Use it effectively by respecting the context window, staying involved via the permission model, and expecting the occasional mistake.