The shortest path from "what is Claude Code?" to "I'm using it."
Languages: English · Русский
A smart assistant that lives in your terminal (the black-and-white text window where developers type commands). You ask it things in plain English – it reads your files, edits them, runs commands, explains errors.
Think of it as ChatGPT that can actually touch your computer.
flowchart LR
You([You]) -- "ask in English" --> Claude["claude<br/>(in your terminal)"]
Claude -- "reads, edits, runs" --> Files[(Your files)]
Files -- "results" --> Claude
Claude -- "answers" --> You
That's the whole experience. You type. It works. You read.
60-second demo video – coming soon. Want to record one? Open a PR linking it here.
Claude doesn't just blurt out edits. Every action runs through a loop:
flowchart LR
Gather["Gather<br/>reads your files,<br/>CLAUDE.md, project"] --> Plan["Propose<br/>shows the change<br/>before doing it"]
Plan -- "you approve" --> Act["Act<br/>edits, runs commands,<br/>creates files"]
Act --> Verify["Verify<br/>checks output,<br/>flags issues"]
Verify -.-> Gather
Three things to know:
- Nothing changes without your "yes." Claude shows the diff and asks before editing.
- Use Plan Mode (
/plan) to see the full plan before any work starts – good for bigger changes. - Read your
CLAUDE.md– the more context that file gives, the less Claude has to ask.
| Your computer | Paste this in your terminal |
|---|---|
| Mac / Linux | curl -fsSL https://claude.ai/install.sh | bash |
| Windows (PowerShell) | irm https://claude.ai/install.ps1 | iex |
Important: Claude Code needs a paid Claude plan (Pro / Max / Team / Enterprise) or a Console account with API credits. The free Claude.ai plan doesn't include it. See pricing.
Prefer no terminal? Get the Desktop app – same agent, regular app window.
Don't have a terminal yet? Start at Step 1 below.
| # | Section | What you'll learn | Time |
|---|---|---|---|
| 1 | Open the terminal | What a terminal is, which one to use | 2 min |
| 2 | Install Claude | Get claude running on Mac or Windows |
2 min |
| 3 | Set up a project | What folders Claude looks at, and how to help it | 3 min |
| 4 | Try real examples | Five things to do today | 5 min |
Bonus: One-page cheatsheet · PDF download
Everything in the four steps above pays off faster if you start from the starter-pack folder:
CLAUDE.md.example— a filled-in brand-voice file. Tweak the placeholders, save asCLAUDE.mdat your repo root, done.prompts/— five battle-tested copy-paste prompts: find feedback themes, blog → social, headline variants, tone-check, localize.
# get just the starter pack
curl -L https://github.com/velesnitski/claude-code-quickstart/archive/refs/heads/main.tar.gz \
| tar -xz --strip=2 claude-code-quickstart-main/starter-packAdapt the brand voice in CLAUDE.md, then use the prompts as-is.
The other excellent guides – Florian Bruniaux's Ultimate Guide, Cranot's Guide, awesome-claude-code – are 10,000+ lines each. Brilliant references, overwhelming for newcomers.
This is the on-ramp. Twelve minutes. Plain language. Pictures. When you're done, graduate:
- Free official courses – Anthropic on Skilljar – start with Claude Code 101, then Introduction to Agent Skills and Introduction to Subagents.
- Long-form references – the guides linked above for exhaustive depth.
- Marketing & content folks explaining Claude Code to others.
- Developers wanting a fast on-ramp before diving into the long guides.
- Team leads onboarding people to AI-assisted workflows.
If you've never opened a terminal before – start at Step 1. We'll walk you through it.
If this guide saved you 5 minutes, give the repo a star – it's the best way to help others find it.
- Translations – open a PR with
README.<lang>.md(Russian already done at README.ru.md). - Screenshots & GIFs – real captures beat ASCII art. Drop them in
docs/images/. - Corrections – fix a section directly, send a PR.
- See CONTRIBUTING.md for details.
MIT – use freely, including in commercial training material.