The free Claude Code course.
A free Claude Code course in five lessons plus a diagnostic, taking you from your first terminal session to subagents that hand off state cleanly. Every lesson is readable without an account. One thing worth knowing before you start: Claude Code itself is not on the free plan - Anthropic ships it with Claude Pro and Max, so the reading here is free and the tool is not.
Prerequisite
You have used Claude in a browser. If not, take Claude 101 first - it is about 70 minutes.
What Claude Code costs.
This is the question people actually arrive with, and most course pages skip it. Anthropic's help centre states that "with Pro and Max plans, you now have access to both Claude on the web, desktop, and mobile apps and Claude Code in your terminal with one unified subscription", and the pricing page lists the free plan's features without Claude Code among them. Both read on 8 August 2026.
- Free plan. Chat, code generation, web search, file uploads, memory across conversations. No Claude Code.
- Pro. $20/month, or $17/month billed annually. Includes Claude Code.
- Team and Enterprise. Also include Claude Code, under their own seat pricing.
- Max. From $100/month, for 5x or 20x the Pro usage.
- Shared limits. Anthropic notes that usage limits are shared across Claude and Claude Code, so terminal work and browser chat draw on the same budget.
Full breakdown of what is and is not free, including the numbers Anthropic does not publish, on the Claude free tier page.
Five lessons, then a check.
Context first, cleverness second. That order is deliberate and it is where most people go wrong.
- 01
What Claude Code is, and what it is not
Time: 10 min
Claude in your terminal, acting on a real codebase, rather than a chat window you copy out of. Anthropic's own Claude Code 101 lesson, condensed to its exam-critical points with citations back to the source.
Open: Claude Code 101 → - 02
Give it context that survives the session
Time: 15 min
CLAUDE.md is the difference between an assistant that re-learns your project every morning and one that does not. The hierarchy matters more than the content: which file wins, and why a repo-level file beats a global one.
Learn: the CLAUDE.md hierarchy → - 03
Hooks - making the loop enforce your rules
Time: 15 min
Where you stop asking the agent to remember a rule and start making the rule mechanical. This is also the single most-tested Claude Code surface on the exam, and the one most people get wrong under time pressure.
Learn: hooks → - 04
Subagents and handing off state
Time: 20 min
One agent with everything in its context is the naive design and it degrades. What actually transfers between a parent and a subagent, what silently does not, and why that boundary is the whole skill.
Learn: subagent state handoff → - 05
When it goes wrong
Time: 15 min
Debugging hooks that do not fire, and the failure modes that look like the model being bad but are actually context being wrong. The difference is most of the job.
Read: debugging Claude Code hooks → - 06
Find out where you stand
Time: 10 min
Ten questions, no account, instant explanations. Claude Code sits inside Domain 3, Agent Operations, which is 20% of the certification - the diagnostic tells you whether that is your weak spot or somebody else's.
Take the free diagnostic →
What you will be able to do.
- Set up a project so Claude Code starts every session already knowing it.
- Decide when a rule belongs in a prompt, in CLAUDE.md, or in a hook.
- Split work across subagents and know exactly what crosses the boundary.
- Tell a context problem from a model problem, which is most of debugging.
- Name your weakest certification domain from your own diagnostic result.
5 questions about the Claude Code course
Every Q is phrased as a real Google search query. Answers cite the same evidence-tagged sources used elsewhere on the site.