July 25, 2026
Claude Code is Anthropic's coding assistant that lives in your terminal instead of a chat window. That description scares off a lot of people who'd otherwise get a lot out of it. You don't need to be a terminal person already. You just need to follow the steps once.
You need Node.js installed first — version 18 or newer. If you don't already have it, grab it from nodejs.org and install it like any other program.
Once that's done, open your terminal (Terminal on Mac, PowerShell or WSL on Windows) and run:
npm install -g @anthropic-ai/claude-code
That downloads Claude Code and makes the claude command available everywhere on your machine. Give it a minute to finish.
Type:
claude
The first time you run it, it'll walk you through logging in with your Anthropic account — either a Claude.ai Pro/Max subscription or an API key, your choice. Follow the prompts, and once it's done you'll land in Claude Code's own prompt, ready to go.
Claude Code works on whatever folder you're in. Before launching it, cd into the project you want help with:
cd path/to/your/project
claude
If you just want to poke around without a real project yet, any empty folder works fine too. It'll read the files that are there and go from there.
Once you're in, just type what you want in plain English:
"Add a dark mode toggle to the settings page."
"Why is this function throwing an error when the input is empty?"
"Set up a basic Express server with one route."
It reads your files, makes changes, and tells you what it did. You can ask it to explain something instead of changing it, ask it to undo a change, or just keep the conversation going like you would with any assistant. It'll ask before doing anything that seems risky, like deleting files or running commands that touch things outside the project.
It works in steps, not all at once. For a big task, it'll often lay out a plan before touching anything, especially if you ask it to. That's a good habit to lean into — for anything nontrivial, ask it to plan first and confirm before it starts writing code.
Slash commands are shortcuts. Type / in the prompt to see a list. /clear starts a fresh conversation, /usage shows you what you've used against your plan's limits (worth checking if you're not sure what those are), /help lists everything else.
It remembers your project, not just your conversation. If you create a file called CLAUDE.md in your project's root folder, Claude Code reads it automatically every time you start a session there. That's the place to put standing instructions — coding conventions, things to always do or avoid, background on the project — so you're not repeating yourself every time.
You can exit anytime. Ctrl+C twice, or just close the terminal. Nothing keeps running in the background after you leave.
Start small. Point it at a real but low-stakes task — fixing a bug, writing a small feature, explaining a file you didn't write — and see how the back-and-forth feels. The learning curve is mostly about trusting it with bigger asks once you've seen it handle smaller ones well.
If you want a sense of how to spend your usage wisely once you're past the basics, I've also written about checking your Claude Code usage limits before you run into one unexpectedly.