Part of my AI & code experiments.
Claude Code is Anthropic’s agentic coding tool that lives right in your terminal. Instead of copying snippets back and forth from a chat window, it reads your project, edits files, runs commands, and works through real tasks with you. Here is exactly how I set it up.
What you need first
- A supported OS: macOS 13+, Windows 10 (1809+), or a modern Linux (Ubuntu 20.04+, Debian 10+).
- 4 GB or more of RAM and a terminal (Bash, Zsh, PowerShell, or CMD).
- A Claude plan that includes Claude Code (Pro, Max, Team, or Enterprise) or an Anthropic Console (API) account for usage-based billing. The free Claude.ai plan does not include Claude Code.
Install Claude Code
The native installer is the simplest option and keeps itself up to date in the background.
macOS, Linux, or WSL:
curl -fsSL https://claude.ai/install.sh | bashWindows (PowerShell):
irm https://claude.ai/install.ps1 | iexPrefer a package manager? Any of these work too:
# Homebrew (macOS/Linux)
brew install --cask claude-code
# npm (needs Node.js 18+)
npm install -g @anthropic-ai/claude-code
# WinGet (Windows)
winget install Anthropic.ClaudeCodeSign in
Open a terminal inside the project you want to work on and start it:
claudeThe first run opens your browser to log in. Sign in with your Claude subscription, or choose the Console/API option if you are billing through Anthropic’s API. After that, Claude Code remembers you.
Verify the install
claude --version
claude doctorclaude doctor checks your installation and configuration and tells you how to fix anything that looks off.
Your first session
From inside a project, just describe what you want in plain English: “add input validation to the signup form”, “explain how auth works in this repo”, or “write tests for the cart service”. Claude Code reads the relevant files, proposes changes, and asks before it edits files or runs commands. You stay in control of every step.
Stay the quality control
This is the part I care about most. The tool writes code fast, but you are still responsible for what ships. Read the diffs, run the tests, and question anything you do not understand before you accept it. Speed without review is how data loss and security holes sneak in: the human is the last line of quality control.
That is the whole setup. Commands and plans change quickly, so check the official Claude Code docs for the latest.
You Are the Quality Control
Claude Code lets you write code faster than ever, but speed without judgement is exactly how bugs, data loss and security holes slip in. My book You Are the Quality Control is a practical guide to building secure software in the age of AI-assisted development, so you can move fast with tools like Claude Code without compromising security, reliability or data safety.
Inside, you’ll learn how to:
- Review AI-generated code with a security-first eye, so risky changes never reach production.
- Put the right foundation, infrastructure and guardrails around AI-assisted projects.
- Build the habits and quality-first mindset that prevent data loss and costly mistakes.
- Stay the human in the loop, because the most important layer of quality control is still you.
