Agentic AI for PHP Developers

Welcome to Agentic AI for PHP Developers — a comprehensive, hands-on series that teaches you to build production-grade AI agents using the claude-php/claude-php-agent framework. This powerful PHP library provides everything you need to create intelligent agents with ReAct loops, tool orchestration, memory management, and advanced agentic patterns.

The claude-php/claude-php-agent framework includes:

• 🔄 Loop Strategies — ReactLoop, PlanExecuteLoop, ReflectionLoop, and StreamingLoop for different task types
• 🛠️ Tool System — Easy tool definition, registration, and execution with JSON schema validation
• 🧠 Memory Management — Persistent state across agent iterations with file-based and in-memory storage
• 🏗️ Agent Patterns — 20+ pre-built patterns including ReAct, Hierarchical, MAKER, and Adaptive agents
• ⚡ Production Ready — Retry logic, error handling, logging, monitoring, and AMPHP-powered async execution

This series is designed for intermediate PHP developers who want to master agentic AI development. You’ll start with the framework fundamentals (agent setup, tool use, loop strategies), then progress to advanced patterns: stateful agents, retrieval pipelines, reflection loops, multi-agent orchestration, and production deployment.

By the end, you will have built a complete agent platform using claude-php/claude-php-agent’s powerful primitives, including custom tool registries, memory systems, RAG pipelines, evaluation harnesses, and multi-agent coordination—all production-ready and ready to power real products.

Want to see an agentic loop in action right now? Here is a simple example using the claude-php/claude-php-agent framework to perform a calculation with tool use:

use ClaudeAgents\Tools\Tool;
$client = new ClaudePhp(apiKey: getenv('ANTHROPIC_API_KEY'));
// Create a simple calculator tool
$calculator = Tool::create('calculate')
    ->description('Perform mathematical calculations')
    ->parameter('expression', 'string', 'Math expression to evaluate')
    ->handler(fn (array $input) => eval("return {$input['expression']};"));
// Create an agent with the tool
$agent = Agent::create($client)
    ->withSystemPrompt('You are a helpful assistant that can perform calculations.');
$result = $agent->run('What is 25 * 17 + 100?');
echo $result->getAnswer(); // 525

Head to Chapter 00 to set up your full environment.

This series is for:

• Intermediate PHP developers who can build web APIs and services
• Laravel/Symfony developers who want to add agentic AI to production apps
• Product engineers building AI features on top of existing services
• Developers already using LLM APIs who want to graduate to agent architectures

You should be comfortable with PHP 8.4+, REST APIs, Composer, and basic async/background job concepts.

Software Requirements:

• PHP 8.4+
• Composer
• Redis (for queues and caching)
• SQLite/MySQL/PostgreSQL (for memory and audit logs)
• Anthropic API Key
• Docker (optional, for deployment chapters)

Time Commitment:

• Estimated total: 35–50 hours
• Per chapter: 60–120 minutes
• Capstone: 4–6 hours

Throughout the series, you’ll build:

1. A production-ready agent runtime using claude-php/claude-php-agent
2. A tool registry with schema validation and permissioning
3. A conversation memory system (short-term + long-term)
4. A RAG pipeline with embeddings + retrieval
5. A planning + reflection loop for multi-step tasks
6. A multi-agent orchestration layer with roles and handoffs
7. A full agentic application shipped with observability and evals

By the end of this series, you will be able to:

• Master the claude-php-agent framework and its agent patterns
• Design agent architectures that go beyond single LLM calls
• Build reliable tool execution with validation, retries, and guardrails
• Implement memory systems and retrieval augmentation in PHP
• Orchestrate multi-agent workflows with specialized roles
• Evaluate agent quality, safety, and cost in production
• Ship and maintain agentic AI applications in real systems

Each chapter explores a core concept of the claude-php-agent framework and shows you how to apply it in production. You’ll start with foundations (agent setup, tool use, loop strategies), build core primitives (tools, memory, plans), then assemble those into full agent workflows. Every part includes implementation guidance, architecture diagrams, and practical exercises.

Ready to start building production-grade AI agents? Head to Chapter 00: Environment Setup and Preparation to install claude-php/claude-php-agent and configure your development environment.