Anthropic · Claude Certified Architect
CCAR-F Free Learning Path
Start with 10 free learning topics for Claude Certified Architect - Foundations (CCAR-F).
10 free/ 30 total
354 free/ 1081 total
Included free
Your free learning route
Follow the learning routeFollow the learning topics in order for the best foundation.
- 01Design prompts with explicit criteria to improve precision and reduce false positivesPrompt Engineering & Structured Output · 1 hr 48 min · 30 questionsLearn now
- 02Apply few-shot prompting to improve output consistency and qualityPrompt Engineering & Structured Output · 2 hr 42 min · 46 questionsLearn now
- 03Design effective tool interfaces with clear descriptions and boundariesTool Design & MCP Integration · 2 hr 24 min · 36 questionsLearn now
- 04Enforce structured output using tool use and JSON schemasPrompt Engineering & Structured Output · 2 hr 52 min · 35 questionsLearn now
- 05Manage conversation context to preserve critical information across long interactionsContext Management & Reliability · 3 hr · 48 questionsLearn now
View all 10 free topics
- 06Design and implement agentic loops for autonomous task executionAgentic Architecture & Orchestration · 1 hr 48 min · 25 questionsLearn now
- 07Implement validation, retry, and feedback loops for extraction qualityPrompt Engineering & Structured Output · 2 hr 24 min · 34 questionsLearn now
- 08Design effective escalation and ambiguity resolution patternsContext Management & Reliability · 2 hr 42 min · 41 questionsLearn now
- 09Implement structured error responses for MCP toolsTool Design & MCP Integration · 2 hr 20 min · 31 questionsLearn now
- 10Manage session state, resumption, and forkingAgentic Architecture & Orchestration · 2 hr 20 min · 28 questionsLearn now
Study resources
CCAR-F Article Guides
Guides and study preparation resources for this certification.
CCAR-F Scope and the Claude Architecture Decision Map
The public scope of Claude Certified Architect - Foundations (CCAR-F) is best understood as an architecture map, not as a collection of isolated product features.
Choosing Single-Shot, Workflow, and Agentic Claude Architectures
The right Claude architecture is usually the least autonomous one that can complete the job safely. Start with a single model call when one bounded response is enough.
Claude Agent Loops, Sessions, and Runtime Controls
An agentic design becomes operable only when its runtime is bounded. In the Claude Agent SDK, an agent loop evaluates the current state, requests tools when needed, incorporates their results, and eventually returns a final result.
Delegation and Subagent Orchestration for Claude Systems
Delegate to a subagent when a focused, independently bounded piece of work would otherwise overload the primary agent's context or mix responsibilities that need different tools and permissions.
Designing Claude Tool Interfaces and Execution Contracts
Claude tool interfaces should be treated as production API contracts, not as a thin list of functions.
MCP Architecture, Primitives, and Trust Boundaries
Model Context Protocol (MCP) is useful when an AI application needs a consistent way to discover and use external context or capabilities. It is not a permission system and it does not make an external system trustworthy by itself.
CCAR-F Video Guides
Explore video explanations for this subject.
8:05CCAR-F Architect Playbook: Designing Claude AI Solutions with the Decision Map
Enterprise AI systems are not built by writing better prompts alone. In this video, we break down the CCAR-F Architect Decision Map and explore how architects approach Claude-powered solutions from a system design perspective. You’ll learn how to move from a business outcome to a production-ready architecture by examining five critical areas: Agentic architecture and orchestration Tool design and MCP integration Claude Code workflows Prompt contracts and structured outputs Context management and
9:46CCAR-F Architecture Patterns: Single-Shot, Workflow, and Agentic Claude Systems
Building an AI agent is not always the right architecture decision. In this video, we break down one of the most important CCAR-F architecture concepts: choosing between single-shot interactions, deterministic workflows, and agentic Claude systems. You’ll learn how architects evaluate: - When a simple model interaction is enough - When application-controlled workflows provide better reliability - When autonomous agent loops are actually justified - How cost, security, and operational risk influe
6:55CCAR-F Claude Agent SDK: Runtime Controls, Sessions, and Permissions Explained
AI agents are powerful, but uncontrolled autonomy creates operational risk. In this video, we explore how architects design bounded Claude Agent SDK runtimes that are secure, predictable, and production-ready. We’ll break down the key runtime concepts behind reliable agent systems: - Agent execution loops and host-controlled boundaries - Session continuation, resumption, and forking - Permission models and approval workflows - Turn limits and budget controls - Context compaction strategies - Tra
9:16CCAR-F Agent Orchestration: Claude Subagents, Permissions, and Delegation Explained
Multi-agent systems are powerful, but adding more agents does not automatically create better AI solutions. In this video, we explore how architects design reliable Claude subagent architectures using clear delegation boundaries, context isolation, permissions, and failure containment. You’ll learn how enterprise AI teams approach: - When subagents are actually necessary - Coordinator-worker architecture patterns - Handoff contracts and context isolation - Least-privilege tool access - Parallel
7:46CCAR-F Claude Tool Use: Designing Secure Tool Interfaces and Execution Contracts
AI agents become production systems only when their tool interfaces are designed with the same discipline as enterprise APIs. In this video, we explore how architects design secure and reliable Claude tool integrations using explicit execution contracts, validation boundaries, and controlled side effects. You’ll learn how to design: - Narrow, outcome-focused tool interfaces - JSON Schema input constraints - Tool execution and result correlation - Safe error handling patterns - Retry and idempote
8:26CCAR-F MCP Architecture: Host, Client, Server Roles and Security Boundaries
MCP is becoming a key pattern for connecting AI systems with enterprise data and capabilities — but protocol connectivity alone does not create security. In this video, we break down the architecture principles behind Model Context Protocol (MCP) and how architects design safe Claude integrations. You’ll learn: - The roles of MCP Host, Client, and Server - How MCP connection lifecycle and capability negotiation work - The difference between Tools, Resources, and Prompts - Where authorization and
PUBLIC BETA OFFER
Save 50% off on an Eligible Learning Plan
Valid for 1-month plan, 3-month plan, 12-month plan · Code LPXBETA50 · Ends Nov 4, 2026