Module 01
12 exercises
Basics
Get set up and productive with Claude fast — the core mental model and first prompts.
3 lessons
Start module
Anthropic · Course
Learn Claude.
Claude is Anthropic's family of AI models, strong at coding, structured writing, and reasoning over long context. This hub breaks Claude down by intent — pick what you are trying to do, then drill into concrete, tested knowledge cards.
Course syllabus
A practical path through Claude.
Move through the modules in order, or return to the part of your work that needs attention now.
Module 02
36 exercises
Coding
Generate, debug, refactor, and architect software with Claude — from a single function to a multi-file codebase.
9 lessons
Start module
Module 03
12 exercises
Writing
Draft, edit, and structure long-form and short-form writing with a consistent voice.
3 lessons
Start module
Module 04
8 exercises
Research
Synthesise sources, compare options, and reason through complex questions.
2 lessons
Start module
Module 05
8 exercises
Workflows
Multi-step systems that chain prompts into a repeatable outcome.
2 lessons
Start module
Module 06
16 exercises
Hidden Features
Capabilities most people miss — Artifacts, Projects, MCP, and Claude Code.
4 lessons
Start module
Module 07
8 exercises
Advanced
Context engineering, tool use, and multi-agent patterns for power users.
2 lessons
Start module
Module 08
8 exercises
Mistakes
The common errors that make Claude output worse — and how to avoid them.
2 lessons
Start module
How to use Claude well
Why people reach for Claude
Claude is Anthropic's family of models, and its reputation is built on three things: disciplined coding, long-form writing that holds structure, and careful reasoning over a large context window. Where a generalist assistant is happy to guess, Claude tends to slow down, ask for the missing piece, and stay faithful to the material you gave it. That makes it a strong default for work where being wrong is expensive — production code, technical documents, contracts, research you will act on.
The practical starting point is the same every time: tell Claude who it is and who you are, then start with the outcome you want rather than the steps. 'You are a senior TypeScript reviewer; I am shipping this to production — flag anything that would break under load' sets a role and a bar in one line. Two other habits pay off immediately — match the model to the task (a fast model for quick edits, a deeper model for architecture and hard reasoning) and show one example of what good looks like, which anchors the output far better than another paragraph of description.
Claude is strongest inside real work, not one-off questions
The long context window is Claude's signature advantage, and it changes how you should use it. Instead of asking narrow questions, give Claude the map before the edit: paste the relevant files, the types, the error, the spec — then ask. For coding specifically, the reliable loop is design before you code, generate one layer at a time, and reproduce a bug before fixing it, so Claude commits to a real theory of the failure instead of a plausible patch.
But a big context window is not a licence to dump everything. The skill that separates good results from noise is context engineering — curating what you include so the signal is not buried. Give Claude the slice that matters, not the entire repository. When you do need it to act on the outside world, describe your tools the way you would brief a new hire and ask it to return structured output you can actually parse.
The features worth learning early
Several Claude features quietly change what is possible. Artifacts turn an answer into a live, iterable document, component, or mini-app you refine in place rather than rebuilding in chat. Projects hold shared context and files across many conversations, so a long engagement stays coherent. Extended thinking lets Claude reason visibly before answering, which is worth turning on for genuinely hard problems. And the Model Context Protocol connects Claude to your own tools and data sources through a standard interface — the doorway to serious, grounded workflows.
If you write code, Claude Code deserves its own mention: let it explore the codebase before it edits, make it plan then wait for approval, and keep a CLAUDE.md that earns its place. Those three habits turn an eager code generator into a dependable teammate.
Avoid the mistakes that waste turns
Most poor Claude sessions trace back to a handful of avoidable errors: vague prompts that force Claude to guess, accepting the first answer without verification, and treating a fluent response as a correct one. Build a verification habit — ask 'what would break this?', request the reasoning, and check claims against the source. Pick the intent below that matches what you are doing and go straight to the tested knowledge cards underneath.
Frequently asked
Is Claude better than ChatGPT?
It depends on the task. Claude is widely preferred for coding discipline, long-form and technical writing, and careful reasoning over long documents, and it tends to stay faithful to the material you provide. ChatGPT is a broader generalist with a deeper consumer feature set and strong multimodal support. Many people keep both and route each job to whichever fits — there is no single winner across the board.
What is Claude best at?
Coding (generation, refactoring, review, and multi-file work), structured long-form and technical writing, and reasoning over a large context window — reading long documents or codebases and answering faithfully. Its large context and careful, low-hallucination style make it a strong choice whenever accuracy and structure matter more than raw speed.
How do I write a good Claude prompt?
Start with the outcome you want, tell Claude who it is and who you are, and show one concrete example of what a good answer looks like. Give it the relevant context — the file, the types, the error — rather than a vague description, and ask it to return structured output when you need to use the result programmatically. Precise role plus a clear target beats a long, unfocused instruction.
What is a context window and why does it matter for Claude?
The context window is how much text — your prompt plus the conversation and any files — Claude can consider at once. Claude's large window lets it reason over long documents and whole codebases in a single pass. The catch is that filling it with irrelevant material buries the signal, so context engineering (curating what you include) matters as much as the size of the window itself.
What is Claude Code?
Claude Code is Anthropic's agentic coding tool that works directly in your terminal and codebase — reading files, planning changes, and editing across a project. It works best when you let it explore before it edits, make it plan and wait for your approval, and keep a focused CLAUDE.md so it understands your project's conventions.
What is MCP (the Model Context Protocol)?
The Model Context Protocol is an open standard for connecting Claude to external tools and data sources — files, databases, APIs, and internal systems — through one consistent interface. It is what turns Claude from a chat window into something that can act on your real data safely and repeatably, and it underpins the more advanced workflows in this hub.
Course checkpoint
Claude AI Master Quiz
Ten questions spanning everything Claude — prompting, coding, writing, research, workflows, hidden features, and the mistakes that quietly waste your turns. Take it as a diagnostic: whatever you miss tells you which topic to read next. No sign-up, instant score, and you can share how you did.