Claude Code: The Complete Free Course (2026)

Claude Code is Anthropic's CLI agent for software engineering. It runs in your terminal, reads your entire codebase, edits files directly, runs your tests, executes shell commands, and commits to git — all driven by plain English. This is the complete pillar guide: what it is, how it works, the full workflow, and links to every deep-dive in the series.

If you read one Claude Code resource, this is it. If you want to go deep, every section links to a dedicated guide.

---

What Is Claude Code?

Claude Code is Anthropic's official CLI coding agent. It runs in your terminal (npm install -g @anthropic-ai/claude-code), reads your entire codebase, edits files directly, runs your tests, and commits to git — all directed in plain English.

The difference from a chatbot is concrete: filesystem access, shell execution, and git integration. You don't copy-paste code between a chat window and your editor. You describe the outcome; the agent ships it and shows you every action it took along the way.

Mental model: a junior engineer with full repo access who never gets tired. This course teaches you to manage that engineer well — because the gap between "has Claude Code installed" and "ships 3x faster with it" is entirely technique.

---

How This Course Works

This page is lesson 1.1 — your orientation. Every lesson is in the course outline on this page (sidebar on desktop, collapsible menu on mobile), with Previous/Next buttons at the end of each one. Lessons are sequenced but self-contained: skip to what you need.

Module 1: Getting Started

Lesson	What you'll be able to do after it
Install, configure, ship your first feature	Working setup + first shipped feature in 30 minutes
The CLAUDE.md guide	Every session starts with full project context instead of zero
Karpathy's CLAUDE.md rules	The 4 rules that cut Claude's mistakes - the config the whole industry copied

If you only do two things: the beginner tutorial and CLAUDE.md. That's 80% of the value for 20% of the time.

Module 2: Workflows & Techniques

Lesson	What you'll be able to do after it
10 prompting patterns	Stop getting generic output - with good/bad examples for each pattern
Hooks	Rules that fire 100% of the time: auto-format, block force-push, quality gates
Sub-agents	3x output with parallel task delegation
Agent Teams	Run a coordinated multi-agent project group
Worktrees	Parallel agents that can't overwrite each other's files
TodoWrite vs Task	Agents that stop forgetting step 5 of 18
Cut API costs 73%	Same output, a third of the bill
Plugins + Skills	Install capabilities instead of re-prompting them
Connect MCP servers	Give the agent your database, GitHub, and internal APIs as tools

The deep-dive on how agents work underneath — function calling, context engineering, memory, sandboxing — lives in the companion AI Agents course. The two courses share a spine; this one is tool mastery, that one is engineering depth.

---

What Does Claude Code Cost?

Two payment routes, one decision rule:

• Claude Pro/Max ($20/$100 per month) — generous usage cap, zero per-token anxiety. Right for solo builders under ~10 hours/week of agent use. Run claude login and you're done.
• Anthropic API key (pay per token) — Sonnet at ~$3/M input, $15/M output. Right for heavy users, teams, and automation. Unoptimized heavy use runs $300-500/month; with the cost playbook (model routing, prompt caching, lean CLAUDE.md, sub-agent discipline), expect $80-150. We cut our own bill 73% with no quality drop.

If you're not sure: start on Pro. You'll know you've outgrown it when you hit the cap twice in a week.

---

Is Claude Code Right for You?

Solo developers — the strongest fit. The agent loop multiplies output without a team. Pro plan covers most weeks.

Small teams (2-10) — shared CLAUDE.md and MCP configs in the repo, individual keys for billing visibility. The multiplier compounds on shared codebases.

Enterprises — API with spend limits, audit logs (one hook away), MCP for internal systems. Production-grade.

Non-developers — honest answer: start with Cursor instead. Claude Code is terminal-first and unforgiving when you don't know what files exist. Come back after a few months of AI-assisted building; the comparison guide covers the decision in full. (Most fast-shipping builders eventually run both: Claude Code for heavy lifting, Cursor for polish.)

---

Start This Week

The skills compound in a specific order: the 30-minute tutorial unlocks CLAUDE.md mastery, CLAUDE.md unlocks workflow discipline, workflows unlock the multi-agent techniques. Skipping ahead works badly; the sequence works well.

So: install Claude Code today, write a CLAUDE.md tomorrow, ship one real thing by Friday. Then come back to Module 2 — the course outline on this page will be here, and so will the structured hands-on version when you want real project walkthroughs.