Symfony AI Documentation

Getting Started

Symfony AI consists of several components and bundles that work together to bring AI capabilities to your application:

• Platform Component: Unified interface to various AI models and providers
• Agent Component: Framework for building AI agents with tools and workflows
• Chat Component: API to interact with agents and store conversation history
• Store Component: Data storage abstraction for vector databases and RAG applications
• AI Bundle: Symfony integration bringing all components together
• MCP Bundle: Integration for the Model Context Protocol SDK

Quick Start

Install the AI Bundle to get started with Symfony AI:

1

$ composer require symfony/ai-bundle

Configure your AI platform:

1
2
3
4
5
6
7
8

# config/packages/ai.yaml
ai:
    platform:
        openai:
            api_key: '%env(OPENAI_API_KEY)%'
    agent:
        default:
            model: 'gpt-4o-mini'

Use the agent in your service:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

namespace App\Agent;

use Symfony\AI\Agent\AgentInterface;
use Symfony\AI\Platform\Message\Message;
use Symfony\AI\Platform\Message\MessageBag;

final readonly class ChatService
{
    public function __construct(
        private AgentInterface $agent,
    ) {
    }

    public function chat(string $userMessage): string
    {
        $messages = new MessageBag(
            Message::forSystem('You are a helpful assistant.'),
            Message::ofUser($userMessage),
        );

        $result = $this->agent->call($messages);

        return $result->getContent();
    }
}

Key Features

Multi-Platform Support

Connect to OpenAI, Anthropic Claude, Google Gemini, Azure OpenAI, AWS Bedrock, Mistral, Ollama, and many more platforms with a unified interface.

Tool Calling

Extend your agents with custom tools to interact with external APIs, databases, and services. Built-in tools available for common tasks.

Retrieval Augmented Generation (RAG)

Build context-aware applications using vector stores with support for ChromaDB, Pinecone, Weaviate, MongoDB Atlas, and more.

Structured Output

Extract structured data from unstructured text using PHP classes or array schemas.

Multi-Modal Support

Process text, images, audio, and other content types with compatible models.

Streaming Support

Stream responses from AI models for real-time user experiences.

Memory Management

Add contextual memory to agents for personalized conversations.

Testing Tools

Mock agents and platforms for reliable testing without external API calls.

Documentation