The Roadmap. 18 Free Courses. 6 Paths. Pick Yours.
Anthropic, the company behind Claude, has been quietly building out a full learning platform. Eighteen free certified courses, taught by the people who built Claude, with verifiable certificates at the end. No subscription required, no credit card required, no upsell at the end.
The platform is moving. Anthropic added five new courses in the last two weeks alone, including one built specifically for small business owners and two more for developers building on top of Claude. By the time you read this, there could be more.
The catch is the same catch with every learning platform: too much content, not enough time. So this guide does one job. It cuts the eighteen courses down to the two or three you actually need, based on what you actually do.
Most readers do not need all eighteen. Some are built for software developers learning to ship apps on top of Claude. Some are for cloud architects deploying Claude inside AWS or Google. A handful, including the eight-hour Building with the Claude API course, will eat a full workday teaching you skills you will never use. Below is the full inventory, color-coded by track, with Rick's picks marked. The next section maps these to roles.
| Course | Track | Duration | |
|---|---|---|---|
| 01 | Claude 101 PICK | Foundation | 1-2 hrs |
| 02 | Introduction to Claude Cowork | Foundation | Not yet estimated |
| 03 | AI Capabilities and Limitations | Foundation | Not yet estimated |
| 04 | AI Fluency: Framework & Foundations PICK | AI Fluency | 2-3 hrs |
| 05 | AI Fluency for Educators | AI Fluency | 2-3 hrs |
| 06 | AI Fluency for Students | AI Fluency | 1-2 hrs |
| 07 | AI Fluency for Nonprofits | AI Fluency | 1-2 hrs |
| 08 | AI Fluency for Small Businesses PICK | AI Fluency | Not yet estimated |
| 09 | Teaching AI Fluency | AI Fluency | 1-2 hrs |
| 10 | Claude Code 101 | Developer | ~1 hr |
| 11 | Claude Code in Action PICK | Developer | ~1 hr |
| 12 | Building with the Claude API | Developer | ~8 hrs |
| 13 | Introduction to Model Context Protocol | Developer | 2-3 hrs |
| 14 | Model Context Protocol: Advanced Topics | Developer | 3-4 hrs |
| 15 | Introduction to Agent Skills | Developer | Not yet estimated |
| 16 | Introduction to Subagents | Developer | Not yet estimated |
| 17 | Claude with Amazon Bedrock | Cloud | 2-3 hrs |
| 18 | Claude with Google Cloud's Vertex AI | Cloud | 2-3 hrs |
Duration estimates are community estimates compiled by Parker Prompts (YouTube, 2026-05-14). Anthropic does not publish official course lengths. Courses added after 5/14 are marked "Not yet estimated." Treat all numbers as ballpark.
Every path in this guide starts the same way: with Claude 101. Take it first regardless of which path you choose. It covers the core features of Claude that almost nobody touches even after using the tool for months, including projects, artifacts, skills, connectors, and research mode. Skipping it leaves you confused later because almost every other course in the academy assumes you already know what those features are.
Claude 101 takes one to two hours, end to end. Walk away with the certificate and you have already done more than most professionals using AI today.
One to two hours. Free. The single course every reader of this guide should take first, no matter which path you pick next.
GO TO CLAUDE 101 →Pick the one that matches your real work, not the one that sounds impressive. Each path lists the courses in order, the estimated time investment, and what you will be able to do after. Time estimates exclude Claude 101 since that is the universal foundation.
You run a team, a business, or a department. You make decisions about how AI gets adopted in your organization. You do not need to write code, but you do need to be the person in the room who actually understands what AI can and cannot do for your people.
After Claude 101 (the universal foundation).
If you only take one thing away from this whole academy, take the Four D's. It is a thinking framework, not a tool tutorial. The difference between leaders who use AI well and leaders who get burned by it is whether they have a framework for deciding what to delegate. This course gives you that framework in three hours.
You own the company. You wear five hats on a normal day. You do not have a tech team, an IT department, or a chief of staff. You need AI to give you back hours every week without you having to become an engineer.
After Claude 101 (the universal foundation).
Anthropic just released this small business course in the last two weeks. The fact that they built a track specifically for owners tells you where the AI conversation is heading. Small business is where the upside lives. You do not have legacy systems, internal politics, or layers of approval. You can deploy a workflow this afternoon and see results by Friday.
You teach. You train. You develop curriculum. You lead a classroom, a workshop, a ministry training program, or a corporate learning function. You need to bring AI into how you teach without losing the rigor of what you teach.
After Claude 101 (the universal foundation).
This course is more useful than people expect. The honest question every educator faces right now is not "should I let students use AI" but "how do I redesign what I teach so AI makes my students sharper, not lazier." This course will not answer it for you, but it will give you the framework to answer it for your own classroom.
You lead a mission-driven organization. You have a small staff, a smaller budget, and a stack of obligations to donors, beneficiaries, and board members. You need AI to multiply your team's output without compromising the values that brought you into this work in the first place.
After Claude 101 (the universal foundation).
Nonprofit budgets do not have room for AI consultants charging Fortune 500 rates. Anthropic built a dedicated track for the nonprofit sector. That is not something tech companies do often. Take it.
You are in school. Undergrad, graduate program, professional certification, seminary. You want to use AI to learn faster and prepare for your career without letting it think for you.
If you already use Claude every day, some of this course will feel basic. Take it anyway, claim the certificate, and use the hour you saved skimming the early modules to actually finish the harder courses on this list. The certificate matters more for you than for anyone else on this guide, because you are still building the LinkedIn that hiring managers will read.
You write code. You are learning to. You want to ship applications, tools, or workflows that run on top of Claude. This path is long, and it is worth it. The combined certificate stack from these courses is one of the most credible AI engineering credentials you can earn for free this year.
After Claude 101 (the universal foundation).
Newly released, no community estimates yet.
Parker Prompts called Claude Code in Action the highest-value course in the entire academy, and he is right. One hour, fifteen lectures, and you walk away knowing how to run agentic coding workflows that actually look like real development work. If you are a developer and you only take one course on this entire guide, take that one. Then come back for the rest when you have a weekend.
Two courses in the academy exist for one reason: enterprise teams deploying Claude inside a specific cloud environment.
If you are an architect on either of those clouds and your company is rolling Claude out at scale, take the one that matches your stack. The lessons translate well to the other side, so if you are evaluating both clouds, taking both gives you a fair comparison.
For everyone else, these are the two safest courses to skip in the entire academy.
Most online AI certificates are noise. They come from third-party course mills run by people who are not building AI themselves. These are different. They come from Anthropic, the company that built Claude. That distinction matters when a recruiter scrolls your LinkedIn.
For every audience on this guide, here is the single course Rick would tell you to take if you only had a couple of hours this month.
Five steps. Do them in this order.
The AI conversation is moving. Six months from now, the people who put in three to five hours this week will be the ones leading rooms where everyone else is still asking what a prompt is.
You have the roadmap. Pick a path. Start this week.