Official study resources
Everything here is published by Anthropic or its official partners. Third-party courseware is deliberately excluded; the exam guides define what is on each exam, and these resources teach it.
Courses
All 21 official courses, with descriptions, enrollment links, and exam relevance, are cataloged in Anthropic courses. The platforms that host them:
| Resource | What it is |
|---|---|
| Anthropic Academy | The public academy: every course free, no partner account needed |
| Anthropic Partner Academy | The certification program's home: prep paths, registration, and partner-exclusive content. Requires a partner sign-in |
| Anthropic on Coursera | Core courses (Building with the Claude API, Claude Code in Action, MCP) delivered through Coursera |
| Build with Claude | Anthropic's index of learning paths for API development |
Documentation
The exams are written against the platform as documented. These are the canonical references:
| Resource | Covers | Most relevant to |
|---|---|---|
| Claude platform documentation | The API: messages, tools, streaming, vision, batches, caching, prompting best practices | Developer, both Architect exams |
| Claude Code documentation | CLAUDE.md, skills, subagents, hooks, headless mode, and best practices | Developer, Architect Foundations |
| Model Context Protocol | The MCP specification, concepts, and SDK guides | Developer, both Architect exams |
| Claude help center | Projects, Artifacts, connectors, and product features | Associate |
Engineering articles
Anthropic's engineering blog covers the exact judgment the scenario questions test. Worth reading in this order:
- Building effective agents: the workflow-versus-agent decision framework that underpins the agent domains
- Effective context engineering for AI agents: context windows, drift, and compaction
- Writing effective tools for agents: tool descriptions and interface design, the substance of the MCP domains
- Equipping agents for the real world with Agent Skills: what Skills are for and when to use them over tools
- Code execution with MCP: efficiency patterns for production MCP use
- Building agents that reach production systems with MCP: integration patterns at production scale
Videos
Anthropic publishes a great deal on its YouTube channel, including the complete AI Fluency course. The official videos page collects the ones that map to exam domains, arranged by exam and playable inline.
The Academy courses are themselves video curricula; for structured watching, they are the primary official video resource.
Code
| Resource | What it is |
|---|---|
| anthropics/courses | Anthropic's own course notebooks: API fundamentals, prompt engineering, evaluations, tool use |
| anthropics/anthropic-cookbook | Runnable recipes for the patterns the exams describe: RAG, tool use, agents, vision |
| anthropics/claude-code | Claude Code itself: issues and releases are a current picture of the tool |
| modelcontextprotocol/servers | Reference MCP servers to read before writing your own |
Matching resources to your exam
- Associate – Foundations: the help center, Prompting 101, and the AI Fluency courses on the public academy.
- Developer – Foundations: platform documentation, the cookbook, the MCP specification, and the API and MCP courses.
- Architect – Foundations: Claude Code documentation and best practices, Building effective agents, Writing effective tools, and the reference MCP servers.
- Architect – Professional: the engineering articles end to end, plus evaluation and RAG material from the cookbook.
Facts last verified against the official sources on 2026-08-01. Repository index