/mem0-integration

name: mem0-integration
description: Mem0 memory layer integration for AI agents. Implement persistent, semantic memory for long-term context retention and personalization.
allowed-tools: Read, Grep, Write, Edit, Bash, Glob, WebFetch
graph:
  domains: [domain:software-engineering]
  specializations: [specialization:ai-agents-conversational]
  skillAreas: [skill-area:context-management, skill-area:retrieval-augmented-generation]
  roles: [role:ml-engineer, role:backend-engineer]
  workflows: [workflow:feature-development, workflow:ml-model-lifecycle]

mem0-integration

Integrate Mem0 (formerly MemGPT) as a universal memory layer for AI agents. Enable persistent memory storage, semantic search across memories, and personalized context retrieval.

Overview

Mem0 provides intelligent memory management for AI applications:

• Persistent storage of conversation history and facts
• Semantic search across stored memories
• User-specific memory isolation
• Automatic memory extraction from conversations
• Support for local and cloud deployments

Capabilities

Memory Operations

• Add memories from text or conversations
• Search memories semantically
• Retrieve relevant context by user/agent
• Update and delete memories
• Get memory history with timestamps

Memory Types

• Conversation memories (dialogue history)
• Fact memories (extracted information)
• Preference memories (user preferences)
• Entity memories (people, places, things)

Storage Backends

• Local SQLite/JSON storage
• PostgreSQL for production
• Qdrant vector database integration
• Cloud-hosted Mem0 platform

Integration Patterns

• LangChain memory integration
• Direct API usage
• MCP server connectivity
• CrewAI and AutoGen compatibility

Usage

Basic Setup

from mem0 import Memory

# Initialize with default local storage
m = Memory()

# Or with custom configuration
config = {
    "vector_store": {
        "provider": "qdrant",
        "config": {
            "host": "localhost",
            "port": 6333,
        }
    },
    "llm": {
        "provider": "openai",
        "config": {
            "model": "gpt-4o-mini",
            "temperature": 0.1,
        }
    }
}
m = Memory.from_config(config)

Adding Memories

# Add memory from conversation
messages = [
    {"role": "user", "content": "I prefer dark mode for all my applications"},
    {"role": "assistant", "content": "I'll remember that you prefer dark mode."}
]
m.add(messages, user_id="user123")

# Add memory from plain text
m.add("User works at Acme Corp as a software engineer", user_id="user123")

# Add with metadata
m.add(
    "Prefers Python over JavaScript",
    user_id="user123",
    metadata={"category": "preferences", "confidence": 0.9}
)

Searching Memories

# Search for relevant memories
results = m.search(
    query="What are the user's preferences?",
    user_id="user123",
    limit=5
)

for memory in results:
    print(f"Memory: {memory['memory']}")
    print(f"Relevance: {memory['score']}")
    print(f"Created: {memory['created_at']}")

Getting All Memories

# Get all memories for a user
all_memories = m.get_all(user_id="user123")

# Filter by metadata
filtered = m.get_all(
    user_id="user123",
    metadata={"category": "preferences"}
)

Memory History

# Get memory changes over time
history = m.history(memory_id="mem_abc123")

for entry in history:
    print(f"Version: {entry['version']}")
    print(f"Content: {entry['memory']}")
    print(f"Updated: {entry['updated_at']}")

LangChain Integration

from langchain_openai import ChatOpenAI
from mem0 import MemoryClient

# Initialize Mem0 client
mem0_client = MemoryClient(api_key="your-api-key")

# Create LLM with memory-enhanced context
llm = ChatOpenAI(model="gpt-4")

def chat_with_memory(user_message: str, user_id: str) -> str:
    # Retrieve relevant memories
    memories = mem0_client.search(user_message, user_id=user_id, limit=5)
    memory_context = "\n".join([m["memory"] for m in memories])

    # Build prompt with memory context
    system_prompt = f"""You are a helpful assistant.

Here is what you remember about this user:
{memory_context}

Use this context to personalize your response."""

    # Generate response
    response = llm.invoke([
        {"role": "system", "content": system_prompt},
        {"role": "user", "content": user_message}
    ])

    # Store new memory from conversation
    mem0_client.add(
        [
            {"role": "user", "content": user_message},
            {"role": "assistant", "content": response.content}
        ],
        user_id=user_id
    )

    return response.content

MCP Server Usage

# Using Mem0 MCP server with Claude
# Configure in claude_desktop_config.json:
{
    "mcpServers": {
        "mem0": {
            "command": "npx",
            "args": ["-y", "@mem0/mcp-server"]
        }
    }
}

Task Definition

const mem0IntegrationTask = defineTask({
  name: 'mem0-integration-setup',
  description: 'Configure Mem0 memory layer for AI agent',

  inputs: {
    storageBackend: { type: 'string', default: 'local' },  // 'local', 'qdrant', 'postgres', 'cloud'
    vectorDimension: { type: 'number', default: 1536 },
    embeddingModel: { type: 'string', default: 'text-embedding-3-small' },
    memoryCategories: { type: 'array', default: ['facts', 'preferences', 'conversations'] },
    userIsolation: { type: 'boolean', default: true }
  },

  outputs: {
    configured: { type: 'boolean' },
    memoryStats: { type: 'object' },
    artifacts: { type: 'array' }
  },

  async run(inputs, taskCtx) {
    return {
      kind: 'skill',
      title: `Configure Mem0 with ${inputs.storageBackend} backend`,
      skill: {
        name: 'mem0-integration',
        context: {
          storageBackend: inputs.storageBackend,
          vectorDimension: inputs.vectorDimension,
          embeddingModel: inputs.embedd