Master Claude AI. Training and certification platform for the Claude Certified Architect by Anthropic.

Claude Certified Architect — FoundationsBETA
No code — learn for freePractitioner — skills & labArchitect — get certified
> _
Start for freeView certification

New to Claude? Free 7-day course →

Passing score

720 / 1000
100PASS →1000
27%Agentic ArchitectureMulti-Agent Systems18%Tool Design & MCP20%Claude Code Config20%Prompt Engineering15%Context ManagementAgent SDK HooksStructured OutputSubagent Spawning27%Agentic ArchitectureMulti-Agent Systems18%Tool Design & MCP20%Claude Code Config20%Prompt Engineering15%Context ManagementAgent SDK HooksStructured OutputSubagent Spawning
ScenarioCustomer Support AgentScenarioCode GenerationScenarioMulti-Agent ResearchDeveloper ProductivityScenarioCI/CD IntegrationScenarioData ExtractionMCP ServersHook SystemScore 720+ScenarioCustomer Support AgentScenarioCode GenerationScenarioMulti-Agent ResearchDeveloper ProductivityScenarioCI/CD IntegrationScenarioData ExtractionMCP ServersHook SystemScore 720+
0Real exam questions
0+Available skills
0Certified domains
0%Free to start

Your entry point

Where do you start?

Free

No code

Learn Claude from scratch

7 free lessons. From your first interaction to real automations. No prior experience needed.

Start for free
Pro

Practitioner

Skills & Prompt Lab

39+ copy-ready techniques. Evaluate and improve your prompts with Claude in real time. From beginner to advanced.

View Pro plan
Architect

Architect

Get certified by Anthropic

65 real questions, 90 minutes, score 100–1000. Prepare to exceed 720 on your first attempt.

Go to simulator

The platform

Everything you need.

Prompt Lab

Pro

Write. Evaluate. Iterate.

Enter a prompt and get instant analysis: score 0–100, strengths, areas for improvement, and an optimized version. Powered by Claude Haiku.

Explore

Skills Library

Partial free

39+ copy-ready techniques.

Slash commands, CLAUDE.md templates, hooks, and MCP configurations. Organized by level: beginner → intermediate → advanced.

Explore

Exam Simulator

Free

Simulator identical to the real exam.

65 questions, 90 minutes, scaled score 100–1000. Exam history, flashcards, and detailed review with per-question explanations.

Explore

News Feed

Free

Updates that matter.

Every Anthropic change analyzed. "Affects the exam" badge so you never miss anything relevant to the certification.

Explore

Free skills

Start using these
right now.

View all 39 skills
slash-commandFREE

/optimize

Refactors selected code to improve performance without changing behavior.

---
name: optimize
description: Refactoriza código para rendimiento
---
Analyze the selected code. Identify bottlenecks.
Refactor for clarity and efficiency.
Preserve all existing behavior.
memoryFREE

CLAUDE.md Base

Project memory template. Claude automatically remembers conventions and commands.

# Proyecto
[Nombre y descripción]

# Stack
- [Framework + versión]
- [Lenguaje + versión]

# Convenciones
- [Naming, style]

# Comandos
npm run dev / build / test
hookFREE

PostToolUse Logger

Hook that logs every tool executed by Claude with a timestamp.

#!/bin/bash
# PostToolUse — log tool calls
echo "[$(date +%H:%M:%S)] $CLAUDE_TOOL_NAME"   >> .claude/tool-log.txt

Claude Certified Architect – Foundations is the official Anthropic certification for professionals who design and deploy systems based on Claude AI. The exam evaluates five domains: Agentic Architecture & Orchestration (27%), Tool Design & MCP Integration (18%), Claude Code Configuration & Workflows (20%), Prompt Engineering & Structured Output (20%), and Context Management & Reliability (15%). Scaled score from 100 to 1,000. Minimum to pass: 720.

The exam

5 domains.
One exam.

01
27% of exam

Agentic Architecture
& Orchestration

Design agentic loops, multi-agent coordinator patterns, subagent spawning, session management, and Agent SDK hooks.

Agentic loopsMulti-agentSubagent spawningSession mgmtAgent SDK

Exam overview

Questions65
Time90 min
Passing score720 / 1000
Scenarios4 of 6 (random)
Go to simulator
02
18%

Tool Design & MCP Integration

03
20%

Claude Code Config & Workflows

Exam format

6 Scenarios

4 picked at random per session

Customer Support Agent
Multi-Agent Research
Code Generation
CI/CD Integration
04
20%

Prompt Engineering & Structured Output

05
15%

Context Management & Reliability

Simulator

See how it works.

Simulator · Domain 1 · Agentic ArchitectureQ 3 / 65

Scenario · Customer Support Agent

You are building a support agent with Claude that must escalate to a human when confidence is low. Which orchestration pattern best describes this architecture?

AChain-of-thought prompting with tool use
BMulti-agent supervisor-worker pattern
CSingle agent with retry loop
DParallel agent fan-out

FAQ

Frequently asked questions.

Start today.
It's free.

No credit card. No commitments. 7 free lessons to get started today.

Create free accountView simulator

Questions? hola@claudepractice.com