Building a Chatbot with Memory

Prerequisites

• Symfony AI Platform component
• Symfony AI Agent component
• OpenAI API key (or any other supported platform)

Step 1: Install Packages

Install the Platform and Agent components via Composer:

1

composer require symfony/ai-platform symfony/ai-agent

Step 2: Create a Memory Provider

The StaticMemoryProvider stores fixed information that should be consistently available to the agent:

1
2
3
4
5
6
7

use Symfony\AI\Agent\Memory\StaticMemoryProvider;

$personalFacts = new StaticMemoryProvider([
    'My name is Wilhelm Tell',
    'I wish to be a swiss national hero',
    'I am struggling with hitting apples but want to be professional with the bow and arrow',
]);

This information is automatically injected into the system prompt, providing the agent with context about the user without cluttering the conversation messages.

Step 3: Add the Memory Input Processor

The MemoryInputProcessor handles the injection of memory content into the agent's context:

1
2
3

use Symfony\AI\Agent\Memory\MemoryInputProcessor;

$memoryProcessor = new MemoryInputProcessor([$personalFacts]);

This processor works alongside other input processors like SystemPromptInputProcessor to build a complete context for the agent.

Step 4: Configure the Agent

The agent is configured with both the system prompt and memory processors. Processors are applied in order, allowing you to build up the context progressively:

1
2
3
4
5
6
7

use Symfony\AI\Agent\Agent;

$agent = new Agent(
    $platform,
    'gpt-4o-mini',
    [$systemPromptProcessor, $memoryProcessor],
);

When a user message is submitted, the MemoryInputProcessor loads relevant facts from the memory provider and prepends them to the system prompt. The agent then generates a personalized response based on both the current message and the remembered context. Because the memory persists across calls, the conversation stays personalized over time.

Step 5: Recall Facts Semantically (Optional)

To recall facts from a large knowledge base instead of a fixed list, swap the static provider for EmbeddingProvider, which retrieves relevant context by semantic similarity:

1
2
3

use Symfony\AI\Agent\Memory\EmbeddingProvider;

$embeddingsMemory = new EmbeddingProvider($platform, $model, $store);

Pass it to the MemoryInputProcessor just like the static provider.

Using the AI Bundle?

If you use the AI Bundle, configure memory declaratively on the agent:

1
2
3
4
5
6
7
8

# config/packages/ai.yaml
ai:
    agent:
        trainer:
            model: 'gpt-4o-mini'
            prompt:
                text: 'Provide short, motivating claims'
            memory: 'You are a professional trainer with personalized advice'

For a dynamic provider, point memory at a service instead. See AI Bundle.

Best Practices

• Keep Static Memory Concise: Only include essential facts to avoid overwhelming the agent
• Separate Concerns: Use the system prompt for behavior, memory for context
• Mind Token Usage: Memory content consumes input tokens, so balance comprehensiveness with cost

Learn More

• Symfony AI - Agent Component - Agent component documentation
• AI Bundle - AI Bundle configuration reference
• Implementing Retrieval Augmented Generation (RAG) - Retrieval Augmented Generation guide