How to set up Claude Code?
To set up Claude Code, install it (`curl -fsSL https://claude.ai/install.sh | bash` on Mac/Linux, `irm https://claude.ai/install.ps1 | iex` on Windows), then run `claude` inside a project folder and sign in with a paid Claude plan or API account. Add a CLAUDE.md file for project rules. Total setup takes about five minutes.
Why — the first-principles explanation
Setup has three logical stages, and each solves a different problem. Install puts the `claude` command on your machine. Authenticate proves who you are so Anthropic can bill your usage—either against a subscription or a metered API account. Configure tells Claude Code about your project so it behaves the way you want.
The reason there's no big settings screen is that Claude Code is designed to work with almost zero configuration. The moment you type `claude` in a folder, it reads the files around it and is ready to help. Everything else—coding standards, which model to use, external tools—is layered on only if you need it.
The most valuable piece of configuration is a CLAUDE.md file at your project root. Claude Code reads it at the start of every session, so it's where you put architecture notes, preferred libraries, and review rules. This works because large language models don't remember past sessions; feeding the same context file each time is how you give Claude a reliable, repeatable briefing instead of re-explaining your project every day.
An example that makes it click
Setting up Claude Code is like hiring a sharp new assistant. First they walk in the door (install). Then they show their badge at the front desk so the company knows to pay them (sign in). Finally you hand them a one-page 'how we do things here' sheet—use these tools, follow this style, check these boxes (that's CLAUDE.md). After that, you just say 'get started,' and they already know the house rules.
How to do it
- Install Claude Code: on macOS/Linux/WSL run `curl -fsSL https://claude.ai/install.sh | bash`; on Windows run `irm https://claude.ai/install.ps1 | iex`.
- Open a terminal in your project: `cd your-project`.
- Start it by typing `claude`, then sign in when prompted using a paid Claude plan (Pro/Max/Team) or an Anthropic Console/API account.
- Try a first task in plain English, for example: `claude "explain what this codebase does"`.
- Optional: create a `CLAUDE.md` file in the project root with coding standards, key commands, and review rules.
- Optional: connect external tools with MCP servers, add hooks, or install the VS Code / JetBrains extension for in-editor use.
Key facts
- Install commands: `curl -fsSL https://claude.ai/install.sh | bash` (Mac/Linux/WSL) or `irm https://claude.ai/install.ps1 | iex` (Windows).
- You must sign in with a paid Claude plan (Pro from $20/month) or an Anthropic Console/API account.
- You start Claude Code by running `claude` inside a project directory.
- A CLAUDE.md file at the project root is read at the start of every session for persistent instructions.
- The same setup (CLAUDE.md, settings, MCP servers) carries across terminal, IDE, desktop, and web surfaces.
Anthropic's agentic coding tool that works in your terminal.
Affiliate link — we may earn a commission at no cost to you.▶ The 60-second explainer (script)
Setting up Claude Code is a three-step job that takes about five minutes. Step one, install it: on Mac or Linux, run the curl install-dot-s-h command; on Windows, the PowerShell irm command. Step two, open a terminal inside your project folder and just type claude. It'll ask you to sign in—use your Claude Pro, Max, or Team plan, or an API account if you'd rather pay per token. That's the minimum; you can now ask it to explain your code or fix a bug in plain English. Step three, and this is the pro move: drop a file called CLAUDE dot M-D in your project root. Write your coding standards, build commands, and review rules there. Claude reads it at the start of every session, so it always knows your house rules. Later you can add MCP tools, hooks, and the VS Code extension, but you're already up and running.
What authoritative sources say
People also ask
Do I need to configure anything before using Claude Code?
No. It works with near-zero config—just install, run `claude`, and sign in. A CLAUDE.md file is optional but recommended.
What is CLAUDE.md?
A markdown file at your project root that Claude Code reads every session. Use it for coding standards, architecture notes, and review checklists.
Can I set up Claude Code inside VS Code?
Yes. Install the Claude Code extension from the Extensions view, though the CLI does the underlying work.
Does my setup follow me between devices?
Your CLAUDE.md, settings, and MCP servers work across terminal, IDE, desktop, and web because they all use the same engine.