What Is Claude Code and Why It Changes Everything
- Understand what Claude Code is and how it differs from chat-based AI tools
- Know what Claude Code can access on your machine by default
- Identify the six environments Claude Code runs in and when to use each
This Is Not a Chatbot
If you have used Claude or ChatGPT in a browser, you already know the workflow: copy some code, paste it in, read the response, copy the answer, paste it back. It works. But it is slow, manual, and breaks the moment your task spans more than one file.
Claude Code is something fundamentally different. It is not a chat interface — it is an agent that lives inside your terminal or IDE and operates directly on your project. It reads your actual files. It runs your actual tests. It commits to your actual git repository. There is no copy-paste loop because Claude Code is already inside your codebase.
What Claude Code Can Actually Touch
Understanding Claude Code's capabilities means understanding what it has access to on your machine. This is the source of its power and the reason you need to approach it thoughtfully:
- Your filesystem — any file in your project directory (and, by default, any file your user account can read)
- Your bash shell — any terminal command: npm, pip, git, curl, database clients, deployment scripts
- Your git history — diffs, branches, commits, remote configuration, and authentication tokens
- Your environment variables — `$DATABASE_URL`, `$AWS_SECRET_ACCESS_KEY`, `$STRIPE_SECRET_KEY`, everything in your shell environment
- The web — Claude Code can make HTTP requests, fetch documentation, and search the internet
- Package managers — it can install npm packages, pip packages, brew formulae, and more
This is the full picture. A later lesson on security will walk you through exactly how to control all of this. For now, the key insight is: Claude Code is not sandboxed by default. It operates with the same permissions your user account has.
The Six Environments
Claude Code runs in six different environments. Each has a different sweet spot:
- Terminal CLI — The core experience. Run
claudein any directory. Best for complex multi-file work, long sessions, automation, and any task that benefits from full bash access. Works on macOS, Linux, and Windows (WSL or native). - VS Code Extension — Integrates Claude Code directly into your editor. Shows diffs inline, lets you reference files with
@, and keeps conversation history in a sidebar. Best for active day-to-day development. - JetBrains IDEs — The same extension experience for IntelliJ IDEA, PyCharm, WebStorm, GoLand, CLion, and others. If you live in a JetBrains IDE, this is your home.
- macOS / Windows Desktop App — A native application with session management, visual diffs, and scheduled background tasks. Good for non-developers or anyone who prefers a GUI.
- claude.ai/code — Runs in Anthropic's cloud with no local setup required. Your code runs on a remote server, not your machine. Good for quick tasks, remote access, or sharing sessions.
- Mobile (Claude iOS) — Limited functionality. Useful for viewing results and sending simple prompts. Not designed for heavy development work.
Most developers start with the terminal CLI and add the IDE extension once they are comfortable. The rest of this course focuses on the CLI because everything you learn there applies everywhere else.
The Real Difference: Multi-File Orchestration
Here is the task that shows clearly why Claude Code is different from a chat interface:
"Rename thegetUserData()function tofetchUser()everywhere in the codebase, update all its callers, update the JSDoc comment, update the tests, and update the README example."
In a chat interface, this task requires you to find every file, copy each one, paste it in, get the answer, paste it back, and hope you did not miss anything. It would take 30 minutes and you would probably miss one file.
In Claude Code: you type the instruction once. Claude searches for all usages with grep, reads every relevant file, makes every edit, runs the tests to verify nothing broke, and reports what it changed. The whole thing takes about 90 seconds.
This is the core value proposition: multi-file orchestration with full codebase context. Claude Code does not see one file at a time — it sees your whole project.
How This Compares to Other AI Tools
| Task | Chat Interface | Claude Code |
|---|---|---|
| Fix a bug across 5 files | Find each file, copy, chat, paste back ×5 | "Fix the auth bug" — Claude reads all 5, edits them |
| Write tests for a function | Copy function → paste → copy tests back | "Write tests for auth.js" — reads file, writes in-place |
| Refactor a module | Manual, file by file | "Refactor fetchUser to async/await" — all files at once |
| Debug a failing test | Read error, find file, copy code, chat | Run tests → Claude reads failure → fixes immediately |
| Rename something everywhere | Find manually, copy each, paste each | One instruction → grep → edit all → verify with tests |
Who This Course Is For
This course is for anyone who writes or manages code — from developers just starting out to experienced engineers who want to work faster. You do not need to be an expert. Claude Code makes beginners productive and experts faster.
By the end of these ten lessons you will know how to install and configure Claude Code, write the CLAUDE.md file that makes it useful for your specific project, control exactly what it can do on your machine, use it safely with real security practices, and build the kind of repeatable workflow that developers use every day in production.
The first step is getting it installed. That is exactly what the next lesson covers.
- Claude Code is an agent that operates directly on your filesystem, shell, and git repo — not a chat interface
- It reads your actual files and runs your actual commands — no copy-paste loop
- The biggest value is multi-file orchestration: rename, refactor, or fix bugs across an entire codebase in one instruction
- It runs in six environments: CLI, VS Code, JetBrains, Desktop app, claude.ai/code, and mobile
- Full filesystem and bash access by default — power and responsibility in equal measure