Claude Code is not a website. It runs in your terminal. A terminal is a text-only window where you type short instructions and the computer types back. No buttons, no menus. Just words going back and forth. It feels different at first. But it is actually simpler than it looks.
If you have never opened a terminal before, here is where to find it. On a Mac, open Finder, go to Applications, then Utilities, then Terminal. On Windows, press the Windows key, type "PowerShell," and open it. That window is your terminal.
You may already use Claude.ai or ChatGPT. Those are web tools. They are useful. But they do not know anything about your actual work. You paste text in, you get text out. Your files stay invisible to them.
Claude Code is different. It runs inside your project folder. It can read your actual files and help you work on them directly. No copy-pasting. No switching tabs.
Here is the simplest way to say it. Claude.ai is like calling a smart friend for advice. You describe the problem in words. Claude Code is like having that friend sit next to you at your desk. They can see your screen, open your files, and read what is actually there.
You need three things before you can install Claude Code.
Node.js. JavaScript usually runs inside web browsers. Node.js is software that lets JavaScript run on your computer instead, outside of any browser. You need it to install Claude Code with npm, which is the method this page uses. Many computers already have it. Check by opening your terminal and typing:
# Check if Node.js is installed
node --version
# You should see something like: v20.11.0
# If you see an error, Node.js is not installed yetIf you see an error, go to nodejs.org and download the LTS version. LTS stands for Long Term Support. It is the stable, recommended one. The installation takes a few minutes and needs no configuration.
Node.js also installs a tool called npm. npm is a package manager, like an app store for developer tools that works by typing commands instead of clicking "Install." You will use npm to install Claude Code.
An Anthropic account. Claude Code on your computer needs to talk to Anthropic's servers to get Claude's responses. There are two ways to give it permission, and you need to pick one before you start.
The first option: a Claude.ai subscription. If you already pay for Claude Pro or Max at claude.ai, you can sign in with that account. Your Claude Code usage counts against your subscription. This is the simplest path for most people.
The second option: an Anthropic API key. An API key is a long string of letters and numbers that acts like a password for a program. It lets Claude Code connect to Anthropic on your behalf. Sign up at console.anthropic.com and create a key under your account settings. With this path, you pay per token. A token is a small chunk of text, roughly part of a word. The cost is billed to your Console account. Keep the key in a password manager. You will only see it once when it is created.
If you are not sure which to pick: choose the Claude.ai subscription path if you already have one. Otherwise, go with the API key.
A folder with some files in it. Any folder works. A coding project is ideal, but even a folder with a few text files is enough to start. You just need something for Claude to look at.
Once Node.js is ready, installation is one command. Open your terminal and type:
# Install Claude Code globally on your computer
# The -g flag means "install globally"
# Without it, the claude command only works in the current folder
npm install -g @anthropic-ai/claude-codeThe installation may take 30 to 60 seconds. The terminal might look stuck for a moment. That is normal. npm is downloading the package. Wait it out.
Once that finishes, you can start Claude Code:
# Start Claude Code for the first time
claudeOn the first run, a browser window opens automatically. Pick your login method: sign in with your Claude.ai account if you have a Pro or Max subscription, or paste your API key from the Anthropic Console.
After that, you are connected. Authentication happens only once. From then on, running claude in any folder opens the tool directly. No login prompt. No key required. It remembers.
Info
Security note: Never paste your API key into a chat window, a document, or anywhere except the Claude Code authentication prompt. Anyone who has your key can use Claude at your expense. This applies to subscription login tokens too. Treat it like a password.
Now navigate to a folder and start Claude Code inside it.
Your computer organizes files into folders. The cd command (short for "change directory") moves you from one folder to another. Claude Code starts working in whatever folder you are currently in. If you start it in the wrong folder, it cannot see the files you want to work on.
# Navigate to your project folder (replace "my-project" with your actual folder name)
cd my-project
# Start Claude Code
claude
# Then type a question, for example:
# > What files are in this folder and what do they seem to be for?Claude reads the files in your folder and responds in plain English. It tells you what it sees, what the files seem to do, and sometimes asks a follow-up question.
That is the key difference. Claude is not guessing from a description you typed. It is reading the actual content, the same way you would if someone handed you a folder and asked you to make sense of it.
You can keep going. Ask it to explain a specific file. Ask it why something looks the way it does. Ask it what you should look at first. The conversation works like any chat, except Claude already has the full picture because it can see your files.
Context is the information Claude has available during a conversation. Think of Claude as a consultant who visits your office for a meeting. During the meeting, they can read everything on your desk. They can answer questions about any of it. When the meeting ends, they forget everything. Next time, they start completely fresh.
In Claude.ai, context is only what you type in the chat. In Claude Code, context also includes your files. That is what makes it useful for real work.
Each time you start Claude Code, the conversation that follows is called a session. It runs from when you type claude until you close it. When you close and open a new session, Claude starts blank again. You will learn how to handle this in the next tutorials, using a file called CLAUDE.md.
Claude Code runs in your terminal and works directly inside your project files. You installed Node.js first, then the claude package with npm. You authenticated once, with either your Claude.ai subscription or your Anthropic API key. Then you ran claude inside a folder and had your first real conversation with AI about your actual work.
Web chat handles general questions well. Claude Code handles work that is right in front of you. That is the foundation everything else builds on.
Next up: 03. IDE Integration: VS Code, JetBrains, and When to Use Each.
Discussion
0 comments· Help other readers by sharing what worked (or didn't) for you.
Sign in to join the discussion.
No comments yet. Be the first to share your thoughts!