Home › Study guides › CCAR-P
CCAR-P study guide · 7 domains · 38 lessons · free, no sign-up
CCAR-P Study Guide: every domain of the Claude Certified Architect - Professional exam, explained
Free study guide for the Claude Certified Architect - Professional (CCAR-P) exam: 38 lessons mapped to the official objectives, with practice.
An independent community resource - not affiliated with or endorsed by Anthropic. The lessons follow the objectives of the official CCAR-P exam guide (Version 1.0, effective July 2026); the practice questions are written from scratch and none come from a real exam.
Start with lesson 1.1 → Read the exam guide first
The exam at a glance
What this guide is
The Claude Certified Architect - Professional exam (CCAR-P) is written against 38 objectives in the official exam guide, grouped into seven domains. This study guide has one lesson for each objective, in the same order, so you can read it front to back or start with the domains that carry the most weight.
Every lesson is built the same way. It starts with a situation an architect actually meets, before any jargon. It then teaches the decision one idea at a time: what the options are, which requirement decides between them, and how you would justify the choice to a security team, an engineering team or an executive. A single fictional organisation runs through the whole lesson, with diagrams, decision tables and short artefacts such as a decision record or a service-level line. It names the traps the exam sets around that objective, because most CCAR-P questions offer several defensible designs and ask for the best one. It ends with a small experiment or design exercise that breaks something on purpose, a list of takeaways, and five practice questions with every answer option explained: four written for the lesson and one from the bank that powers the timed simulators.
The guide assumes you know systems design. It does not assume you have run Claude in production for long: every Claude-specific term is explained the first time it appears. Plan on about 20 minutes per lesson.
How to use it
- Read the CCAR-P exam guide page once. It has the facts (63 questions, 120 minutes, pass at 720 on a 100 to 1,000 scale) and the registration rules.
- Start with the heavy domains. Integration is almost a fifth of the exam, and Solution Design and Evaluation are about a sixth each; Developer Productivity is 7%.
- Do not skip Domain 6. Stakeholder communication and lifecycle management is 14% of the exam, as much as governance, and it is the domain engineers most often leave unprepared.
- When you have read every domain, sit a 63-question timed simulator in the question bank. Treat a score under 75% as "not yet", whatever the real pass mark is.
Course outline
7 domains, 38 lessons, about 14 hours of reading. Domains are in the order the official guide lists them; the weight beside each is the share of the exam it carries.
Domain 1: Solution Design & Architecture
Designing Claude solutions: from business problem to architecture, choosing workflow or agent, multi-agent design and business value. 17% of CCAR-P.
- 1.1 Turning a business problem into a Claude solution
- 1.2 End-to-end architecture: input, processing, output and feedback loops
- 1.3 Augmented LLM, workflow or agent: choosing the pattern
- 1.4 Multi-agent systems: when to split the work and how to orchestrate it
- 1.5 Decomposition: cutting complex work into steps you can test
- 1.6 Aligning a Claude solution to business value pillars
Domain 2: Claude Models, Prompting & Context Engineering
Choosing a Claude model, system prompts and guardrails, prompting techniques, context and token use, and reuse with caching and Skills. 13% of CCAR-P.
- 2.1 Selecting Claude models: trade-offs, constraints and the evidence to decide
- 2.2 System prompts, templates and prompt guardrails
- 2.3 Zero-shot, few-shot and chain-of-thought: choosing the technique per decision
- 2.4 Optimising the context window: select, compact and measure tokens
- 2.5 Prompt reuse: caching, modular prompts and Agent Skills
Domain 3: Integration
Connecting Claude to enterprise systems: capability bloat, auth, accuracy vs latency, observability, RAG, retrieval and protocols. 19% of CCAR-P.
- 3.1 Capability bloat: auditing an agent's tools and permissions
- 3.2 Authentication and authorisation: whose identity does each tool call use?
- 3.3 Accuracy versus latency: measure both, then justify the configuration
- 3.4 Observability at scale: tracing, sampling and scoring LLM traffic
- 3.5 Designing a RAG pipeline: chunking, metadata and indexing
- 3.6 Retrieval strategies matched to the data and the question
- 3.7 MCP, API, CLI or agent-to-agent: choosing the integration mechanism
- 3.8 Progressive discovery or monolithic context: what an agent loads up front
Domain 4: Evaluation, Testing & Optimization
Measuring and improving a Claude solution: metrics, evaluation datasets, A/B tests, diagnosing failures, cost and latency, and monitoring. 16% of CCAR-P.
- 4.1 Evaluation metrics: defining accuracy, latency, cost, safety and security
- 4.2 Evaluation datasets and test frameworks: mixed methods that hold up
- 4.3 A/B testing and iterative improvement: proving a change on live traffic
- 4.4 Diagnosing system issues: prompt failures, hallucinations and model mismatch
- 4.5 Optimising tokens, latency and cost without losing quality
- 4.6 Monitoring a live Claude system: logs, dashboards, alerts and incidents
Domain 5: Governance, Safety & Risk Management
Governing Claude solutions: guardrails, failure modes, human review, GDPR, HIPAA and FedRAMP, and bias, fairness and transparency. 14% of CCAR-P.
- 5.1 Guardrails in layers: input, model, output and action controls
- 5.2 Risks, limitations and failure modes: from threat model to risk register
- 5.3 Human-in-the-loop validation: putting review where the risk is
- 5.4 Compliance by design: GDPR, HIPAA and FedRAMP on Claude
- 5.5 Bias, fairness and transparency when Claude helps decide about people
Domain 6: Stakeholder Communication & Lifecycle Management
The people side: structured discovery, communicating decisions and trade-offs, expectations and SLAs, documentation and the lifecycle. 14% of CCAR-P.
- 6.1 Structured discovery: stakeholders, evidence and requirements design can use
- 6.2 Communicating architecture decisions: records, audiences and honest trade-offs
- 6.3 Expectations, SLAs and feedback loops for an LLM system
- 6.4 Documenting the architecture and guiding the team that builds it
- 6.5 Supporting the lifecycle: phase gates, handoff and owned iteration
Domain 7: Developer Productivity & Operational Enablement
Setting teams up with Claude: configuring Claude Code for a team, AI-assisted development workflows, and debugging and operations support. 7% of CCAR-P.
Three things that surprise people
It is a decision exam. Few questions ask what a parameter does. Most describe a production system with a problem and ask which change is best, and several options would work. The right one fixes the cause and fits the stated requirement; the others monitor the problem, compensate for it or over-build.
A third of it is not about technology. Governance, stakeholder communication and lifecycle management together are 28% of the exam. Discovery, service levels, decision records and human review are tested as architecture work, not as soft skills.
Some questions have more than one answer. Items can be multiple-response, and each item says how many options to select. Read that line before the options.
Frequently asked questions
Is this study guide really free?
Yes. Every lesson, every practice question on these pages, and the full question bank are free. The bank asks for an account only because the quiz engine has to store your answers to score them. Nothing here is behind a paywall.
Do I need CCAR-F first?
No. The official guide lists no prerequisites. CCAR-F covers the mechanics of agents and Claude Code in more depth; CCAR-P assumes you can design, justify and govern a whole solution.
Is it enough on its own to pass CCAR-P?
It covers every objective in the official guide, which is what the questions are written against. It cannot replace experience: the guide recommends having built and operated at least one end-to-end Claude solution, including retrieval, evaluation and observability.
Is this affiliated with Anthropic?
No. It is an independent community resource. The exam facts come from Anthropic's published exam guide; the lessons and questions are written from scratch and none come from a real exam.
189 CCAR-P practice questions, free
189 practice questions here: timed practice sets and a full 63-question exam simulator on a 120-minute timer. Every answer option carries its own explanation, so a lucky guess reads the same as a wrong answer.
Open the CCAR-P question bank → Exam facts, rules and registration →
The question bank is free. It asks for an account only because the quiz engine has to store answers to score them and show which domains are weak. The questions on this page need nothing.