On August 20, 2026, Anthropic opened Claude Academy, a learning site for picking up AI skills in a structured way. It lives at academy.claude.com and brings together 289 resources across courses, tutorials, and use cases. Everything is free, and the material can be browsed without creating an account. The site pulls together product-specific guidance for Claude, development work with the API and MCP, and the fundamentals of AI itself.
Separate Entry Points for Each Product
The front page starts with product-based entry points. There are 5 of them: Claude.ai for thinking through problems, drafting documents, and analyzing data in conversation; Claude Cowork for handing off whole tasks and getting finished deliverables back; Claude Code for building and debugging in the terminal or an IDE; Claude Tag for bringing Claude into team channels and assigning it work; and Claude Platform for wiring Claude into your own products through the API and MCP.
The catalog holds 22 courses spanning beginner material through production-level work. Claude 101 runs 13 lessons over 2.5 hours, covering first conversations, effective prompting, Projects, Artifacts, Skills, and connected tools. Claude Code 101 is shorter at 12 lessons and 1 hour, focused on how Claude Code works and the core workflows around it. Introduction to Claude Cowork takes 14 lessons and 2.5 hours to cover workspace setup, supplying context, and delegating multi-step work.
The developer track is a different weight class. Building with the Claude API spans 67 lessons and 8 quizzes over 9 hours, running through prompting, tool use, RAG, agents, MCP, and production patterns. Separate cloud-specific courses exist for Amazon Bedrock (65 lessons, 8 hours) and Google Cloud's Vertex AI (66 lessons, 8.5 hours). MCP gets both an introduction (10 lessons, 1 hour) and an advanced course covering sampling, notifications, and roots (11 lessons, 1.5 hours).
The 4D AI Fluency Framework Sits at the Center
The backbone running through the material is what Anthropic calls AI Fluency. The 4D framework stands for Delegation, Description, Discernment, and Diligence, and it treats deciding what to hand to AI, conveying intent, evaluating what comes back, and using the results responsibly as one continuous skill set.
The anchor course, AI Fluency: Framework & Foundations, runs 14 lessons over 4 hours. Alongside it sits AI Capabilities and Limitations (13 lessons, 3.5 hours), which covers how large language models predict the next token and where knowledge, working memory, steerability, and context length run out. Shorter pieces such as the 7-minute tutorial The 4 Properties of AI offer a way in when time is tight.
Audience-specific variants are a notable feature. Students, educators, pK-12 teachers, nonprofits, small businesses, and builders each get a version of the same 4D framework translated into their own setting. The pK-12 track even includes a Train the Trainer course so teachers can run AI fluency workshops for their colleagues.
Saved Progress and Completion Badges
Signing in is not required to read the material, but signing in with a Claude account saves learning progress. Eligible courses award a completion badge after a final assessment, and the site recommends courses based on stated interests and past enrollments.
Looking at the breakdown, 22 items are courses and most of the remainder are use cases and tutorials. These include practical recipes such as producing a one-page account brief in 10 minutes for sales calls, or sorting account health into green, yellow, and red with 2 moves for the week. That gives both the person settling in to study and the person who only wants today's task finished somewhere to land. Short courses on newer concepts such as agent skills and subagents are in there as well.
Summary
Claude Academy bundles 3 directions into one free site: product-specific use of Claude, development with the API and MCP, and AI fundamentals. It offers 22 courses within a total of 289 resources, with the 4D framework as a common structure running from students through to developers. The emphasis on how to collaborate with AI, rather than on feature documentation, is what sets it apart from the usual pile of docs.
