Beginner lesson plan
How to use Claude Code: a six-lesson course for beginners
Updated November 17, 202610 min read
How to use Claude Code comes down to six skills: install it, chat with it inside your project folder, give it a notes file, make it plan before big jobs, undo its mistakes, and learn a short list of commands. This page teaches each skill as a small lesson you can finish in one evening. All you need is a terminal, a project folder and a paid Claude plan or an API account.
After these six lessons you can
- Install Claude Code on Mac, Linux or Windows and sign in.
- Ask questions about any project and get a first change made safely.
- Write a short CLAUDE.md so Claude remembers your rules every session.
- Use plan mode to see the approach before any file changes.
- Roll back a bad change with
EscEscor/rewind. - Use the dozen commands that cover almost every day of work.
Before lesson 1: what to have ready
Claude Code is Anthropic's AI coding agent. The main version is a CLI (a program you control by typing into a terminal window). It reads the files in your project, changes them, runs commands and uses git (the tool that keeps a saved history of your project) for you. You describe the goal in normal English and approve what it does.
- A terminal. On a Mac that is the Terminal app. On Windows it is PowerShell. Anthropic has a short terminal guide if you have never opened one.
- A project folder. An existing project is best, but an empty folder works for a first try.
- An account that includes Claude Code. As of November 2026 that means Claude Pro, Max, Ultra, Team or Enterprise, or a Claude Console account with pre-paid API credits. The Free plan does not include it. Our Claude Code pricing page compares the options.
The CLI is not the only way in. The same agent also runs on the web, as a desktop app, inside VS Code and JetBrains editors, in Slack and in GitHub Actions. This course sticks to the terminal because every other version builds on the same ideas.
Lesson 1: install the Claude Code CLI
Open a terminal and paste the one line that matches your computer. These are Anthropic's native installers as of November 2026.
Mac, Linux or WSL
curl -fsSL https://claude.ai/install.sh | bashWindows PowerShell
irm https://claude.ai/install.ps1 | iexWindows Command Prompt (CMD)
curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmdThe installer shows no progress bar while it downloads, so give it a minute. When it finishes, close the terminal, open a fresh one and type claude --version. A version number means it worked. If the shell says claude is not recognised, the install folder is not on your PATH (the list of places your computer looks for programs) yet, and Anthropic's troubleshooting page walks through the fix.
Tip
On Windows, install Git for Windows too. Claude Code then uses its Bash tool. Without it, Claude Code falls back to PowerShell. Our Claude Code on Windows page covers the Windows details.
Homebrew (brew install --cask claude-code) and WinGet (winget install Anthropic.ClaudeCode) also work, but those copies do not update themselves. The native install updates in the background, and claude update forces an update whenever you like.
Lesson 2: your first chat with Claude Code
- 1
Go to your project
In the terminal, type
cdfollowed by the path to your project folder, then press Enter. - 2
Start Claude Code
Type
claudeand press Enter. The first time, it asks you to log in. - 3
Sign in once
Your browser opens. Pick your Claude subscription or your Console account. Claude Code stores the login, so you will not need to do this again. Use
/loginlater if you want to switch accounts. - 4
Ask about the project
Before changing anything, ask questions. Claude reads the files it needs by itself, so you never have to paste code in.
what does this project do?
what technologies does this project use?
where is the main entry point?
explain the folder structureNow ask for a small change, such as add input validation to the user registration form. Claude finds the right file and shows you the edit. If it asks before changing the file, choose Yes to approve. To point Claude at one exact file, type @ and start typing the file name.
How often Claude asks first depends on the permission mode (a setting for how much it may do without asking). As of November 2026, new terminal sessions start in auto mode where it is available, and in it a background safety check approves most actions for you. Press Shift+Tab to cycle through the modes. From auto mode, one press takes you to manual mode. While you learn, manual mode is the calmer choice, because Claude asks before every edit and command and you see each step.
Lesson 3: give Claude a CLAUDE.md notes file
Every new Claude Code session starts with a blank memory. CLAUDE.md is a plain text file that Claude reads at the start of every session, so it is where you write down the things Claude cannot guess from the code: how to run the app, how to run the tests, and your house rules.
- 1
Generate a starter
Type
/init. Claude looks through the project and writes a CLAUDE.md with the build commands, test commands and habits it found. If one already exists, it suggests improvements instead of overwriting it. - 2
Cut it down
Read the file and delete anything obvious. Anthropic's advice is to keep each CLAUDE.md under 200 lines, because longer files use more context and get followed less closely.
- 3
Add what only you know
The command that starts the dev server (the program that runs your app on your own computer while you work), folders Claude must not touch, the way you name things.
- 4
Edit it later
Type
/memoryany time to open your CLAUDE.md files and change them.
# My project notes for Claude
- Start the app: npm run dev
- Run tests: npm test (run them before saying a task is done)
- Never edit files in /legacy
- Use the existing Button component, do not add new UI libraries./CLAUDE.md(or./.claude/CLAUDE.md): shared rules for this project. Commit it so teammates get them.~/.claude/CLAUDE.md: your personal rules for every project on this computer../CLAUDE.local.md: your private notes for one project. Add it to.gitignoreso it stays off GitHub.
A CLAUDE.md can pull in another file with @path/to/file. Remember that CLAUDE.md is advice that Claude follows almost always. For a rule that must never be broken, a Claude Code hook is the stronger tool.
Lesson 4: make Claude plan before it builds
For anything bigger than a small fix, let Claude research first and show you a written plan. In plan mode, Claude can read files and look around, but it does not edit your code until you approve. You can start a plan in one line:
/plan add a contact form that emails me. use the email helper we already have. done means I can send a test message.- 1
Read the plan
Check which files it names, what it will delete and how it will test the result.
- 2
Push back
Choose No, keep planning and tell Claude what to change, or press
Ctrl+Gto edit the plan text yourself. - 3
Approve
Pick Yes, manually approve edits while you are learning, so you see each change as it happens.
You can also press Shift+Tab until the bar at the bottom says ⏸ plan mode on. Our Claude Code plan mode guide shows the same task done with and without a plan.
Lesson 5: undo mistakes with checkpoints
Claude Code quietly saves a checkpoint (a snapshot of your files) before each prompt you send. If a change goes wrong, you can step back to any earlier prompt in the session. Checkpoints even survive closing and resuming a session.
- 1
Stop early
Press
Esconce to stop Claude in the middle of what it is doing. The work so far is kept. - 2
Open the rewind menu
Press
Esctwice on an empty prompt, or type/rewind(also called/undo). - 3
Pick the moment
The menu lists every prompt you sent. Choose the last one before things went wrong.
- 4
Pick what to restore
Choose Restore code and conversation, Restore code (keep the chat), or Restore conversation (keep the files). Summarize from here shrinks a long chat without touching files.
Watch out
Rewind only tracks files that Claude changed with its own editing tools. A file that only a shell command changed, or an edit made by a subagent, may not come back. Keep committing to git as your real safety net.
Claude Code keeps file snapshots for the 100 most recent checkpoints in a session and clears old ones about 30 days after a session was last used, as of November 2026.
Lesson 6: the commands you will type every day
Claude Code has two kinds of commands. Shell commands start it from the terminal. Slash commands run inside a session and start with /. Type / on its own to see every command you have, and use Tab to finish a name.
claude # start a chat in this folder
claude "fix the build error" # start with a first request
claude -p "explain this file" # answer once, then exit
claude -c # continue your last chat here
claude -r # pick an older chat to resume
claude update # get the newest versionInside a session, the habit that matters most is /clear between unrelated tasks. A fresh chat is cheaper and gives better answers than one that carries an hour of old, unrelated history. When you want to keep the thread but trim it, /compact summarises it instead. To leave, type /exit or press Ctrl+D twice.
Claude Code commands cheat sheet
| Command | Where | What it does |
|---|---|---|
claude | Terminal | Start an interactive session |
claude -p "query" | Terminal | One answer, then exit (good for scripts) |
claude -c / claude -r | Terminal | Continue the last chat / choose one to resume |
claude -w name | Terminal | Start in its own git worktree (worktrees guide) |
/init | Session | Write a starter CLAUDE.md |
/memory | Session | Edit CLAUDE.md files and auto memory |
/clear | Session | Start a fresh conversation |
/compact | Session | Summarise the chat to free space |
/context | Session | Show what is filling the context window |
/plan | Session | Enter plan mode, optionally with a task |
/model, /effort | Session | Switch model, or how hard Claude thinks |
/usage | Session | Session cost and plan limits (/cost is an alias) |
/rewind | Session | Roll back code, chat or both |
/resume | Session | Reopen an earlier conversation |
/code-review | Session | Check your changes for bugs (review guide) |
/mcp, /hooks | Session | Manage connected tools and automations |
/help | Session | List everything available |
Your first project with Claude Code, start to finish
Reading about commands only goes so far. Here is a small first project that uses every lesson once. It builds a one-page to-do list website in an empty folder.
- Make a new folder called
todo-app, open a terminal there and rungit initso you have save points. - Run
claudeand type/plan build a single-page to-do list in plain HTML, CSS and JavaScript. tasks should stay after a page refresh. keep it to three files. - Read the plan. Ask
what could go wrong with this plan?and approve when you are happy. - When the files exist, ask
how do I open this in my browser?and try it. - Ask for one change at a time, such as
add a button that clears finished tasks. Check the result after each one. - If a change breaks the page, press
Esctwice and restore the code from before that prompt. - Run
/initso the next session knows how the project works, then askcommit my changes with a clear message.
Tip
Once one project works, many people start a second and third session for other projects and lose track of which terminal is waiting. VibeCodingDeck lays out up to 12 Claude Code sessions in one window on Windows 10/11 or Apple-silicon Macs, each with a status light, using the Claude Code and subscription you already have.
Where the official Claude Code docs live
Anthropic's Claude Code docs are at code.claude.com/docs. After this course, the most useful pages are the CLI reference for every flag, the commands list for every slash command, and best practices for habits. You can also ask Claude Code about itself, for example how do I create custom skills in Claude Code?. Anthropic also runs a free self-paced course, Claude Code 101, on Anthropic Academy.
For the next step after the basics, read our Claude Code tips page, then learn subagents when you want Claude to hand side jobs to helpers.
Beginner questions about Claude Code
Is Claude Code hard to learn if I have never coded?
+
The basics are not hard. You type requests in plain English. The new skills are opening a terminal, starting claude in the right folder, and a handful of commands like /clear, /plan and /rewind. Reading Claude's plans carefully matters more than knowing how to code.
Can I use Claude Code without the terminal?
+
Yes. Anthropic also offers Claude Code on the web, a desktop app, and extensions for VS Code and JetBrains. Third-party apps such as VibeCodingDeck put the CLI inside a visual window.
What is the difference between claude and claude -p?
+
claude opens a session you can keep chatting in. claude -p "query" answers once and exits, which suits scripts and quick one-off questions.
How do I update Claude Code?
+
Native installs update themselves in the background, and claude update updates on demand. Homebrew users run brew upgrade claude-code, and WinGet users run winget upgrade Anthropic.ClaudeCode.
How do I get Claude Code to remember my project between sessions?
+
Put the important facts in CLAUDE.md. Run /init to create one, keep it short, and edit it with /memory. Claude reads it at the start of every session.
Sources
Run up to 12 AI coding agents in one window.
Claude Code, Codex, Gemini, Grok and more, side by side. Windows and Mac. From $29 a month.
Get startedKeep reading
Before and after
Claude Code plan mode: one task, with and without a plan
Cost calculator
Claude Code pricing: plans, API rates and which one fits you
24 numbered tips
Claude Code tips and tricks: 24 habits that pay off
Windows install guide
Claude Code Windows install: choose your method, then get it running
Buyer's map
Claude Code GUI: a buyer's map of every visual way to run Claude Code