Installing Claude Code
Before you can build with Claude, you need it installed and a subscription that gets you access.
This chapter is the install. The next two chapters cover a markdown viewer you’ll want alongside it and the configuration that makes Claude actually usable in practice.
Pick a plan
Anthropic offers a few entry points:
- Claude Pro: the basic individual plan. Good if you’re exploring: some daily use in the web app, modest Claude Code work, limited heavy sessions.
- Claude Max: higher-limit individual plan. The right choice if you’re using Claude Code as your primary working surface all day. Multiple tiers depending on how heavily you use it.
- Claude Team / Enterprise: for small teams and organizations. Centralized billing, admin controls, SSO on the Enterprise tier.
- API (pay-as-you-go): separate from the consumer plans. Only needed if you’re building against the Anthropic API directly.
For people working through this course, Max is usually the right plan if you’re in Claude Code every day. Pro works if you’re just exploring. The API plan is orthogonal: you can have either (or both).
Current pricing at claude.com/pricing.
Set up your terminal
On macOS, the built-in Terminal works fine: it’s just an access point to your machine, the same way a browser is an access point to the web. But you’ll want something with better tab and pane ergonomics. I recommend iTerm2.
Install from iterm2.com or via Homebrew: brew install --cask iterm2.
Why iTerm over Terminal:
- Tabs with
Cmd-T: just like a browser. - Split panes: right-click inside a pane and choose “Split Vertically” or “Split Horizontally.” Have Claude running in one pane while you watch logs, edit files, or run commands in another.
- Sane defaults and power-user features: better colors, better search.
If you already prefer a different terminal (Alacritty, WezTerm, Warp, Ghostty, etc.), use it. The only real requirements are tabs and splits.
Install Claude Code
Claude Code is the command-line harness: the local tool that runs in your terminal, reads your repo, and calls the model. It’s the surface most of this course assumes.
Canonical install instructions are in the Claude Code setup docs. On macOS, two common paths:
- Native installer: prefer this. It auto-updates itself as Anthropic ships new versions.
- Homebrew: also works, but updates only when you run
brew upgrade. Easy to fall behind model or feature releases.
Either way, you can uninstall and reinstall at any time without losing anything important. Session logs, CLAUDE.md, and Skills live in your home directory and your repos, not inside the install.
After installing:
claude
The first run walks you through authentication. Sign in with the same account you used for your Claude plan; Claude Code will use that subscription.
Your first session
Once claude runs, do a quick smoke test:
- Go to a repo on your machine (any repo).
- Run
claude. - Ask something concrete: “describe the architecture of this project” or “what files handle authentication?”
- Watch it use
glob,grep, andreadto answer.
Claude has no memory of your project yet, but it can explore it on the fly.
Before you get too deep, head to Configuring Claude and turn on auto mode. Without it, every session is a stop-and-go experience. With it, Claude flows.
Keep the official docs handy
Install commands, plan details, and feature lists change. Bookmark:
- Anthropic pricing
- Claude Code docs
- Anthropic status page: when something seems broken