Free Claude certification course
Claude certification course lessons are free here: 13 mirrors of Anthropic's Skilljar material plus 10 focused deep dives. Each entry maps the source material to the certification exams and links to the relevant concepts and production scenarios, so you can study in order or jump to the domain your diagnostic flagged.
Use the Claude certification comparison to choose the credential, follow the path to becoming a Claude Certified Architect, then confirm your readiness with the Claude Certified Architect mock exam.
Start here · D1 mirrorSubagents: Context Management + Task Delegation4 lessons · ~25 min · the D1 primitive the exam weights mostSubagents: Context Management + Task Delegation
Specialized helpers spawned into their own context window that return only a summary.
Claude 101: Foundations for Everyday Work
What Claude is, how to write prompts that land, and how to organize work using Projects, Artifacts and Skills.
Claude Code 101: Workflows + Context Mastery
The daily-driver course for the agentic coding tool that reads your codebase, edits files, and runs commands.
Introduction to Claude Cowork: Delegation for Knowledge Work
Claude Cowork is the same agentic architecture that powers Claude Code, retargeted at knowledge work: analysis, research, writing, and the documents you produce every day. The shift the course teaches is from conversation to delegation: describe an outcome, review the plan, let it run, and review the finished file. It's the right mental model for any exam question about when to use chat vs an agent.
Claude Code in Action: Integration Patterns
The eight integration surfaces that turn Claude Code into a real engineering tool.
AI Fluency: The 4D Framework for Working with AI
AI Fluency teaches a four-competency framework for collaborating with AI: Delegation (deciding what to give the AI), Description (communicating clearly), Discernment (evaluating outputs), and Diligence (taking responsibility). The course is light on technical depth but heavy on the disciplined judgment that separates effective AI users from prompt-and-pray users. Treat it as the operating manual under every other Skilljar course you will take.
Building with the Claude API: Foundations to Agents
The comprehensive bottom-up tour of the Anthropic SDK.
MCP Foundations: Tools, Resources, Prompts
An open spec that lets any Claude client talk to any external system through a thin server.
MCP Advanced: Sampling, Notifications, Roots, Transports
The capabilities production servers actually need.
Agent Skills: Reusable Prompts in Claude Code
Reusable markdown files that teach Claude Code how to handle specific tasks automatically.
AI Capabilities and Limitations: A Mental Model of the Machine
Generative AI has four properties that each sit on a continuum from capability to limitation: Next Token Prediction, Knowledge, Working Memory, and Steerability. Most real-world failures are two properties colliding (a hallucinated citation is Next Token Prediction meeting a Knowledge gap), and naming the pair points you straight to the fix. This is the durable mental model that stays useful even as models keep improving.
Claude with Google Cloud Vertex AI: Deployment + GCP Integration
This 93-lesson course teaches the same Claude API surface as claude-api-foundations but accessed through Google Cloud's Vertex AI rather than the direct Anthropic API. The deployment-specific substance is the AnthropicVertex SDK, gcloud Application Default Credentials auth, Model Garden enablement, project + region binding, and the regional model-availability model. Everything else (prompt engineering, evals, tool use, RAG, MCP, agents) mirrors Course 6 lesson-for-lesson.
Claude in Amazon Bedrock: Deployment + AWS Integration
This 83-lesson course teaches the same Claude API surface as claude-api-foundations but accessed through Amazon Bedrock rather than the direct Anthropic API. The deployment-specific substance is the AWS boto3 SDK, IAM-bound auth, the Bedrock converse API shape, regional model availability with cross-region inference profiles, and Bedrock-only features (Guardrails, Knowledge Bases, Agents). Everything else (prompt engineering, evals, tool use, RAG, MCP) mirrors Course 6 lesson-for-lesson.
Is There an MCP Certification? What Actually Certifies Model Context Protocol Skill
There is no standalone MCP certification. Model Context Protocol competence is certified inside Anthropic's CCA-F, specifically Domain 2 (Tool Design + MCP Integration, 18%). Here is what that domain tests and how to prep for it.
Agentic AI Certification: Which One Proves You Can Build Agents on Claude
Several agentic AI certifications exist. For building agents on Claude, the credential is Anthropic's CCA-F, where Domain 1 (Agentic Architectures, 27%) is the single heaviest domain. Here is the landscape and how to prep.
Why Subagents Don't Automatically Inherit CLAUDE.md and How to Fix It
Subagents spin up with empty context - CLAUDE.md is a parent-session filesystem read, not a transmitted artifact. Here is how to force inheritance so subagents respect your project conventions on every call.
Designing MCP Error Contracts That Trigger Claude's Retry Behavior
Claude only retries MCP tool calls when the error contract tells it the failure is transient. Here is the JSON envelope, the four required properties, and the exam trap that swallows errors inside 200 OK responses.
Stop MCP Tool Descriptions Eating Your Context
MCP tool descriptions can silently eat 25% of your context window. Audit your token budget, apply progressive disclosure, prune redundant schema fields, and inject only task-relevant tool subsets per turn.
Architecture-Aware, Refactor-Authorized Workflows: Getting Useful Results from AI Coding Agents
Pass the agent an explicit architecture contract, scope its refactor authorization, force a plan-confirm-execute loop, inject retrieval over codebase embeddings, and validate output against an architecture linter in CI.
Debugging Claude Code Hooks That Fire Silently
Four silent-failure causes, the canonical debug loop with --debug, the registered-but-skipped vs erroring-silently vs never-registered taxonomy, --hooks-dry-run, and the CCA-F anti-pattern of confusing missing output with successful execution.
How to Ensure Claude Uses the Latest User-Provided Context After a Sequence of Updates
Five techniques for keeping Claude anchored on the most recent context after multiple updates - recency placement, override anchors, version stamps, compression, and uptake validation.
Handling Token Limits When Processing Large JSON Responses in an MCP Client/Server Flow
Four canonical patterns for shrinking oversized MCP tool results - field projection, pagination, chunk-with-summarize, and streaming - mapped to CCA-F D2/D5 with before/after token counts.
How to Handle Very Long Prompts for Complex Business Contexts Without Splitting Them
Five techniques for keeping a long business-context prompt coherent without breaking it into smaller calls - front-loaded decisions, hierarchical structure, XML sectioning, scoped ignores, and the CCA-F D4 trade-off.
8 questions about the free course
Answers focus on the lessons, study order, time and exam usefulness.
Are the Anthropic Academy courses really free?
Yes. All 13 courses on Skilljar are free and open to anyone - no Partner Network membership, no paid Anthropic account required. Just an email signup with Skilljar. The CCA-F certification exam itself does require Partner Network access, but the prep material is universally free.
Can I pass CCA-F just by watching the Anthropic Academy courses?
Maybe, but you'll be on the edge. The courses cover concepts thoroughly but don't drill the production-scenario combinations the exam tests. Beta cohort data: candidates relying only on Academy averaged 680/1000; those who added scenario drills averaged 750/1000.
Is there a recommended order for the courses?
Yes. (1) Claude 101. (2) Claude Code 101. (3) MCP Foundations + MCP Advanced. (4) AI Fluency Framework. (5) Subagents, Skills, Hooks. (6) Platform courses (Bedrock, Vertex) only if you target those platforms in your work.
Is there a free Claude certification course?
Yes. The course hub is free and combines Anthropic Skilljar lesson mirrors, focused deep dives, concept lessons, production scenarios and links to practice. You can begin without paying for a course and move directly to the domain your diagnostic identifies.
How is the Claude certification course organized?
Use the Skilljar mirrors for Anthropic's lesson sequence, the deep dives for focused technical questions, the concept pages for individual primitives and the scenarios for complete production architectures. Each entry links to the next relevant study or practice step.
What is the Anthropic Academy?
Anthropic's free Skilljar-hosted course platform covering Claude foundations, Claude Code, MCP, AI Fluency, and platform-specific deployments (Bedrock, Vertex). 13 courses total as of May 2026, all free, no Anthropic Partner Network membership required. The Knowledge pillar here distills each course to the exam-critical points.
Which Anthropic Academy course is most exam-relevant?
Claude 101 (foundational, covers ~30% of CCA-F D4 and D5 content), Claude Code 101 (D2 tool design + D3 ops), and MCP Foundations (D2 integration). The AI Fluency Framework course is dense but covers the 4D framework concept that appears as a single high-density question in most exam attempts. Bedrock and Vertex are platform-specific - only critical if your Practice mock shows D5 weakness.
How long does each course take?
Claude 101 is the longest at ~4 hours of video content. Most other courses are 60-120 minutes. The Knowledge summaries on this site condense each to a 10-15 minute read covering the exam-critical 20% - useful as a primer before watching, or as a refresher after.