Claude Academy & Documentation
Anthropic's free learning hub: Claude Academy's courses, from Claude 101 to building agents, plus the developer documentation. Hands-on, current, and written for real work.
My take
Claude Academy is amazing. I look at it frequently and recommend it to all my coworkers. If you use Claude, this is where you start. What you won't find there anymore is a separate prompt engineering course, and I think that's right: a prompt should be written in the context of the whole system around it.
At a Glance
Pros
- + Claude Academy covers everything from a first conversation to Claude Code, Cowork and agents
- + Free, and kept up to date with the product
- + Courses for non-technical people as well as developers
- + Tool use and agent documentation is the clearest in the industry
Cons
- − Naturally focused on Claude — you won't learn GPT-specific patterns here
- − The developer courses and docs assume some programming comfort
- − Courses and documentation live on two separate sites
Best for: Anyone who uses Claude and wants to go from 'it works sometimes' to 'I know why it works'
Full review
Anthropic doesn’t just build Claude and hope you figure it out. Its courses and documentation are some of the best learning material in AI right now, and they’re completely free. They now live in two places: the courses in Claude Academy, the developer documentation at platform.claude.com/docs.
Claude Academy
This is the highlight. Claude Academy has around 30 courses and keeps adding more. For everyday users there’s Claude 101, an introduction to Cowork and Anthropic’s AI Fluency courses, with versions for different audiences. For builders there’s Claude Code 101, Claude Code in action, courses on the Model Context Protocol, skills and subagents, and a nine-hour course on building with the Claude API.
What makes it better than most AI course material: it teaches you how the pieces fit together, not just what to type. That matches how I think about AI skills: understanding beats memorizing. I look at it often and recommend it to everyone at work.
Where prompt engineering went
Anthropic used to have a dedicated prompt engineering course. The tutorial still exists on GitHub, but it was written for the Claude 3 models, and the docs now point to their prompting best practices as the living reference. I think that’s the right call. Prompting isn’t a skill you learn in isolation anymore: a good prompt is written in the context of the whole system around it, the knowledge Claude has access to, the tools it can use and the job it’s part of.
Documentation that’s actually readable
Anthropic’s docs read more like a well-written tutorial than a typical technical reference. The guides on tool use, system prompts and building agents are clear enough that non-developers can follow the concepts, even if they won’t be writing code themselves.
The cookbook has practical examples you can adapt, from summarization to data extraction to building simple agents.
Who should start here
If Claude is your main AI tool, start here before watching YouTube videos or buying courses. For non-technical professionals: Claude 101, the Cowork introduction and the AI Fluency course. Developers should go on to Claude Code 101, the API course and the docs.
The one caveat: everything is Claude-specific. That’s fine if Claude is your daily driver, but pair this with broader resources if you want to understand the full AI toolkit.
Added: 2026-03-20 · Last updated: 2026-10-22