You’ve heard people talking about Claude Code. Maybe a developer friend mentioned it. Maybe you saw it trending on LinkedIn or Reddit. Maybe you searched for “AI coding tools” and this name kept coming up. Whatever brought you here, you’re in the right place — because this is the page that explains everything, from scratch, with no assumed knowledge.
Claude Code is one of the most powerful AI tools available right now. It doesn’t just chat about code — it actually does things inside your computer. It reads your files, writes code, runs commands, and works through multi-step tasks on its own. That’s a fundamentally different kind of AI tool, and getting your head around what it is (and how to use it) opens up a whole new way of working. This guide takes you from zero to your first productive session — step by step.
## 📋 TL;DR
– Claude Code is an AI coding agent from Anthropic that runs in your terminal and can read, write, and edit your files.
– It’s different from Claude.ai (the browser chatbot) — it acts on your computer, not just in a chat window.
– You need a paid Claude subscription (starting at $20/month on Pro) to use it. There’s no free tier.
– Installation is straightforward: Anthropic now offers native installers for Mac, Windows, and Linux that require no extra software.
– You talk to Claude Code in plain English — no special commands required to get started.
– The best first step is to point it at a small, real project and ask it to explain the code or fix one specific problem.
– This guide links to detailed articles on every sub-topic so you can go deeper wherever you need.
Table of Contents
- What Exactly Is Claude Code?
- How Claude Code Is Different from Claude.ai
- What You Need Before You Install
- How to Install Claude Code
- Claude Code Pricing — What It Costs
- Your First Session — What to Expect
- What Claude Code Can (and Can’t) Do
- Common Beginner Mistakes and How to Fix Them
- Quick Reference Cheatsheet
- What to Learn Next
- Key Takeaways
What Exactly Is Claude Code?
Claude Code is an AI coding agent built by Anthropic — the company that also makes Claude, the AI assistant. But Claude Code isn’t a chatbot you open in a browser tab. It’s a tool that runs in your terminal (also called the command line or command prompt — the text-based window where developers type instructions to their computer).
Here’s the clearest way to picture it: imagine hiring a very capable developer who sits right inside your computer. You tell them what you want in plain English. They read your project files, make changes, run tests, and report back. That’s Claude Code. It doesn’t just suggest code for you to copy-paste — it can actually open files, write to them, and run commands on your machine.
The word “agent” is important here. An agent doesn’t just answer one question and wait. It takes a goal, breaks it into steps, and works through those steps until the job is done. Ask Claude Code to “add a login page to my app” and it might read five files, write three new ones, update a config file, and run your test suite — all in one go.
Claude Code launched as a limited preview in February 2025 and became generally available in May 2025. Since then it has become one of the most widely used AI developer tools in the world. Both experienced engineers and people who have never written a line of code use it every day.
💡 Pro Tip: Claude Code works with nearly any programming language or project type. You don’t need to be building something specific — you can point it at an existing website, a Python script, a mobile app, or even a folder of text files and it will figure out the context.
How Claude Code Is Different from Claude.ai
If you’ve already used Claude.ai — Anthropic’s browser-based chat interface — you might wonder why Claude Code even exists. It’s a fair question. Both tools use the same underlying AI models. But they work in completely different ways.
Claude.ai lives in your browser. You type a message, Claude types back. It’s a conversation. It has no access to your files, your computer, or your projects. If you paste in some code and ask for help, it can explain it and suggest changes — but you have to copy those changes out and apply them yourself. It’s a chat partner.
Claude Code lives in your terminal. It has direct access to your file system — meaning it can open, read, and write real files on your computer. It can run commands (like running your tests, installing packages, or starting a server). It can hold an entire codebase in context at once, understanding how all your files relate to each other, not just the snippet you paste in.
Think of it this way: Claude.ai is like calling a consultant on the phone. They give great advice, but you have to go implement it yourself. Claude Code is like that consultant sitting next to you at your desk, with their hands on the keyboard.
There’s also a practical difference in where you use them. Claude.ai is great for questions, drafting, brainstorming, and general tasks. Claude Code is purpose-built for working inside actual projects. Many people use both — they’ll brainstorm an architecture in Claude.ai, then switch to Claude Code to build it.
[Claude Code vs Claude.ai — What’s the Difference?]
What You Need Before You Install
Before you install Claude Code, let’s check the basics. Getting these right first saves a lot of frustration.
A paid Claude subscription or API access. There is no free version of Claude Code. The cheapest entry point is the Claude Pro plan at $20 per month, which includes Claude Code. If you don’t have a subscription yet, you’ll need to create an Anthropic account at claude.ai and subscribe before Claude Code will work. (More on pricing below.)
A supported operating system. Claude Code runs on:
- macOS** — any recent version works well
- Windows** — supported via the native Windows installer or Windows Subsystem for Linux (WSL)
- Linux** — fully supported with a native installer
A terminal. On Mac, this is the built-in Terminal app (or a third-party option like iTerm2). On Windows, this is PowerShell, Command Prompt, or Windows Terminal. On Linux, whatever terminal emulator you use. If you’ve never opened a terminal before, don’t panic — the installation guides linked at the end of this article walk you through it step by step.
That’s it. You don’t need to be a developer. You don’t need to know how to code. You don’t even need Node.js installed anymore — Anthropic’s native installers bundle everything Claude Code needs. This is a recent change (2026) that made setup significantly simpler than older guides describe.
[Claude Code System Requirements]
💡 Pro Tip: If you find older blog posts or YouTube tutorials telling you to run
npm install -g @anthropic-ai/claude-code, those are outdated. Anthropic now recommends the native installer, which requires zero dependencies. Skip npm unless you have a specific reason to use it.
How to Install Claude Code
Installation is a single command in your terminal. Anthropic provides native installers for each operating system that handle everything automatically — including keeping Claude Code updated in the background.
On Mac and Linux, open your terminal and run the installer script Anthropic provides on their official documentation page. The script downloads Claude Code, places it in the right location on your system, and adds it to your PATH (which is just the list of places your computer looks when you type a command — so that typing claude in your terminal will actually find Claude Code).
On Windows, Anthropic provides a PowerShell installer command that does the same thing. If you’re a Windows developer who uses WSL (Windows Subsystem for Linux), you can also run the Linux installer inside your WSL environment.
After installation, confirm everything worked by running:
claude --version
You should see a version number printed. If you do, Claude Code is installed.
Logging in happens the first time you run claude in a project directory. A browser window will open for an OAuth sign-in flow — you log in with the Anthropic account that has your Claude subscription, and Claude Code stores a session token locally. You won’t need to log in every time after that.
[How to Install Claude Code on Mac]
[How to Install Claude Code on Windows]
[How to Install Claude Code on Linux]
claude --version command with a successful version number response, and then the browser login screen that appears on first runClaude Code Pricing — What It Costs
Claude Code isn’t priced as a separate standalone product. It’s included in Anthropic’s Claude subscription plans — so the bill you pay for Claude also gives you Claude Code access.
Here’s how the tiers break down:
Claude Pro — $20/month (or $17/month billed annually). This is the starting point for individual users. You get Claude Code in the terminal, the desktop app, and the web interface, plus access to Sonnet and Opus models. The usage allowance is enough for focused, daily coding sessions. Most beginners start here.
Claude Max 5x — $100/month. Five times the usage allowance of Pro, plus priority access during busy periods. This tier makes sense if you find yourself hitting Pro limits regularly.
Claude Max 20x — $200/month. Twenty times Pro’s allowance. For people who use Claude Code as their primary development tool for most of the working day, this tier is significantly cheaper than paying per-token through the API.
API pay-as-you-go. Instead of a subscription, you can pay directly per token (roughly 4 characters of text). This gives you no monthly ceiling, but costs vary based on usage. For light or irregular use, this can work out cheaper than a subscription. For daily use, the subscription plans usually win.
Team and Enterprise plans also exist for organizations, with centralized billing and admin controls.
One important note: there is no free tier for Claude Code. The free Claude account does not include it.
According to Anthropic’s own data, 90% of Claude Code users stay below $12 per day in equivalent API usage. For most beginners, the $20/month Pro plan is the right starting point.
[Claude Code Pricing — Free vs Pro vs API Explained]
Your First Session — What to Expect
You’ve installed Claude Code, you’re logged in, and you’re staring at your terminal. Now what?
Start by navigating to a project folder. In your terminal, use the cd command (cd stands for “change directory” — it’s how you move between folders in a terminal) to get to a folder with some files in it. This could be a code project, a folder of scripts, even a directory with text files.
cd my-project-folder
Then launch Claude Code:
claude
Claude Code will start up and display a prompt, waiting for you to type. You’re now in a session — a conversation with Claude Code that has full awareness of all the files in your current directory.
Talk to it like you’d talk to a person. Try asking it to explain what’s in your project:
What does this project do? Give me a plain English summary.
Or ask it to fix something specific:
The login button isn't working. Can you find the problem and fix it?
Claude Code will read your files, think through the problem, and then tell you what it’s going to do before it does it. It’ll ask for your permission before making changes to files. You can say yes, review what it proposes, or ask it to try a different approach.
To end a session, type /exit or press Ctrl+C.
The most important thing to understand: Claude Code learns the context of your project as you talk to it. The longer your session, the better it understands what you’re building. Don’t start a new session for every question — keep your conversation going and it gets smarter about your specific project as you go.
[Your First 10 Minutes with Claude Code]
[What is the Claude Code CLI?]
What Claude Code Can (and Can’t) Do
Knowing what Claude Code is actually capable of helps you use it well — and stops you wasting time asking it for things it’s not built for.
What it can do:
- Read and understand your entire codebase, including files across many folders
- Write new code from scratch based on your description
- Edit existing code to fix bugs, add features, or refactor for clarity
- Run shell commands (like installing packages, running tests, or starting a server)
- Work with Git — the version control system most developers use — including making commits and reading your project history
- Explain what existing code does in plain language
- Generate documentation, write tests, and review code for problems
- Work through multi-step tasks autonomously, without you guiding each individual action
- Handle context windows up to 1 million tokens, meaning it can hold enormous codebases in memory
What it can’t do:
- Browse the internet in real-time by default (though this can be configured)
- Access files outside the directory you started it in, unless you give it permission
- Remember previous sessions — each new
claudesession starts fresh (though you can use a file calledCLAUDE.mdin your project to give it persistent context about your project) - Guarantee correct output — like any AI, it can make mistakes. Always review changes before you commit or deploy them.
The CLAUDE.md file deserves a special mention. It’s a plain text file you put in the root of your project that tells Claude Code important context: what the project does, what coding conventions you use, which files are off-limits, and so on. Claude Code reads it automatically at the start of every session. Creating one is one of the highest-leverage things you can do as a Claude Code user.
[What is Claude Code?]
Common Beginner Mistakes and How to Fix Them
Mistake 1: Giving Claude Code a vague task
The problem: You type something like “fix my project” or “make this better.” Claude Code doesn’t know what’s broken, what “better” means to you, or where to start.
The fix: Be specific. Name the exact file, describe the exact problem, and explain the expected behavior. “The submit button in ContactForm.jsx throws a TypeError when the email field is empty. It should show a validation message instead” gets dramatically better results than “fix the contact form.”
Mistake 2: Starting a new session for every question
The problem: Every time you start a new session with claude, Claude Code loses all context about your project. Beginners often quit and restart constantly, which means Claude Code never builds up a useful understanding of what they’re building.
The fix: Stay in one session as long as you’re working on the same project. Claude Code gets more useful the more it knows about your codebase. Only start a new session when you’re genuinely switching to a different project or task.
Mistake 3: Not reviewing changes before accepting them
The problem: Claude Code asks if it can make a change, you say yes without reading what it’s about to do, and now your file is different in a way you didn’t expect.
The fix: Always read what Claude Code proposes before approving it. Claude Code will show you a diff (a side-by-side view of what changed) before making any edits. Take 30 seconds to read it. If something looks wrong, ask Claude Code to explain its reasoning or try a different approach.
Mistake 4: Using an outdated installation method
The problem: You find a 2024 blog post or YouTube tutorial that tells you to run npm install -g @anthropic-ai/claude-code. You run it, hit errors, get frustrated.
The fix: Use Anthropic’s native installer, which doesn’t require Node.js and is the officially recommended path as of 2026. Check Anthropic’s official documentation for the current installation command — it’s one line.
[How to Update Claude Code]
[Claude Code Login Problems — How to Fix Them]
Mistake 5: Expecting Claude Code to be perfect
The problem: Claude Code writes some code, it doesn’t work, you assume the tool is broken or not worth using.
The fix: AI makes mistakes. Claude Code makes mistakes. The workflow isn’t “ask once, deploy forever” — it’s “ask, review, test, iterate.” When Claude Code’s output doesn’t work, tell it what happened. Paste the error message. Ask it to try again. It learns from the feedback within a session and usually gets it right on the second or third attempt. Treat it like a capable colleague who sometimes needs correction, not an oracle.
Quick Reference Cheatsheet
| **Task** | **What to do** |
|---|---|
| Install Claude Code | Run Anthropic’s native installer (see [LINK: How to Install Claude Code on Mac/Windows/Linux]) |
| Check your version | `claude –version` in your terminal |
| Start a session | Navigate to your project folder, then run `claude` |
| End a session | Type `/exit` or press `Ctrl+C` |
| Ask for help | Type your question in plain English — no special syntax needed |
| Give project context | Create a `CLAUDE.md` file in your project root |
| See available commands | Type `/help` inside a Claude Code session |
| Switch AI models | Type `/model` inside a session to choose between Sonnet, Opus, etc. |
| Check usage cost | Type `/cost` inside a session to see your token usage |
| Minimum plan required | Claude Pro ($20/month) — no free tier exists |
| Supported platforms | macOS, Windows, Linux |
| Context window | Up to 1 million tokens |
| Need to log in again | Delete your session token and run `claude` — the browser login flow will appear |
What to Learn Next
This page gives you the foundation. These articles go deeper on every piece of the puzzle — work through them in whatever order matches where you’re stuck or curious.
Understanding Claude Code:
- [What is Claude Code?] — A deeper look at how Claude Code works and why it’s built differently from other AI tools.
- [What is the Claude Code CLI?] — Everything you need to know about the CLI (command-line interface) and how sessions work.
- [Claude Code vs Claude.ai — What’s the Difference?] — Side-by-side comparison if you already use Claude in your browser.
- [Claude Code vs Cursor vs GitHub Copilot] — How Claude Code stacks up against the other major AI coding tools.
Getting Installed:
- [How to Install Claude Code on Mac] — Step-by-step Mac installation with screenshots.
- [How to Install Claude Code on Windows] — Step-by-step Windows installation, including WSL setup.
- [How to Install Claude Code on Linux] — Step-by-step Linux installation for Ubuntu, Debian, and other distros.
- [Claude Code System Requirements] — The full technical specs: what hardware and OS versions are supported.
Money and Account:
- [Claude Code Pricing — Free vs Pro vs API Explained] — A plain-English breakdown of every plan and which one is right for you.
Troubleshooting:
- [Claude Code Login Problems — How to Fix Them] — If you can’t log in or your session keeps expiring, start here.
- [How to Update Claude Code] — How to keep Claude Code current (the native installer handles this automatically, but it’s worth knowing how).
Your First Day:
- [Your First 10 Minutes with Claude Code] — A guided walkthrough of your very first session, with prompts to try.
Key Takeaways
- Claude Code is an AI agent, not a chatbot.** It acts inside your computer — reading files, writing code, running commands — rather than just answering questions in a browser window.
- You don’t need to be a developer to use it**, but you do need to be comfortable opening a terminal and typing basic commands. The installation articles on this site walk you through that step by step.
- There is no free tier.** Claude Code requires at least a Claude Pro subscription ($20/month). That plan is enough for most beginners and casual users.
- Installation is easy in 2026.** Anthropic’s native installer handles everything — no Node.js or other dependencies required. Ignore older guides that say otherwise.
- The learning curve is mostly about prompting.** Claude Code itself is straightforward. The skill worth developing is how to describe your tasks clearly and specifically. The more precise your instructions, the better your results.
Last updated: July 2026. Claude Code is under active development — if you notice anything that’s changed, check Anthropic’s official documentation at docs.claude.com for the latest details.