MemMachine Client

A Python client library for the MemMachine memory system.

Features

• Simple API: Easy-to-use interface
• Memory Management: Add and search episodic and profile memories
• Context Awareness: Automatic context retrieval for better responses
• Error Handling: Robust error handling with retry mechanisms
• Type Safety: Full type hints for better development experience

Installation

pip install memmachine-client

Development installation

If you are working on the package locally inside the MemMachine monorepo, install it from source instead:

pip install -e packages/client

Quick Start

Basic Usage

from memmachine_client import MemMachineClient

# Initialize client
client = MemMachineClient(
    base_url="http://localhost:8080",
    timeout=30
)

# Create a memory instance
memory = client.memory(
    group_id="my_group",
    agent_id="my_agent",
    user_id="user123",
    session_id="session456"
)

# Add memories
memory.add("I like pizza", metadata={"type": "preference"})
memory.add("I work as a software engineer", metadata={"type": "fact"})

# Search memories
results = memory.search("What do I like to eat?")
print(results)

API Reference

MemMachineClient

The main client class for interacting with MemMachine.

Constructor

MemMachineClient(
    api_key: Optional[str] = None,
    base_url: str = "http://localhost:8080",
    timeout: int = 30,
    max_retries: int = 3,
    **kwargs
)

Methods

• memory(group_id, agent_id, user_id, session_id) - Create a Memory instance
• health_check() - Check server health

Memory

Interface for managing episodic and profile memory.

Methods

• add(content, producer, produced_for, episode_type, metadata) - Add a memory
• search(query, limit, filter_dict) - Search memories
• get_context() - Get current context

Examples

Basic Memory Operations

from memmachine_client import MemMachineClient

client = MemMachineClient(base_url="http://localhost:8080")

# Create memory instance
memory = client.memory(
    group_id="demo_group",
    agent_id="demo_agent",
    user_id="user123",
    session_id="demo_session"
)

# Add memories with metadata
memory.add("I like pizza", metadata={"type": "preference", "category": "food"})
memory.add("I work as a software engineer", metadata={"type": "fact", "category": "work"})

# Search memories — Memory.search() returns a SearchResult Pydantic model
results = memory.search("What do I like to eat?")
print(f"Episodic memory: {results.content.episodic_memory}")
print(f"Semantic memory: {results.content.semantic_memory}")

# Search with filters (user metadata fields require the `m.` / `metadata.` prefix)
work_results = memory.search("Tell me about work", filter_dict={"m.category": "work"})
print(f"Work results: {work_results.model_dump()}")

Multiple Users

from memmachine_client import MemMachineClient

client = MemMachineClient(base_url="http://localhost:8080")

# Create memory instances for multiple users
users = ["alice", "bob", "charlie"]
memories = {}

for user in users:
    memories[user] = client.memory(
        group_id="team_group",
        agent_id="team_agent",
        user_id=user
    )

# Add user-specific memories
memories["alice"].add("I'm a frontend developer", metadata={"role": "frontend"})
memories["bob"].add("I'm a backend developer", metadata={"role": "backend"})
memories["charlie"].add("I'm a DevOps engineer", metadata={"role": "devops"})

# Search across users
for user, memory in memories.items():
    results = memory.search("What is your role?")
    print(f"{user}: {results}")

Error Handling

from memmachine_client import MemMachineClient

try:
    client = MemMachineClient(base_url="http://localhost:8080")

    # Check server health
    health = client.health_check()
    print(f"Server health: {health}")

    # Create memory instance
    memory = client.memory(
        group_id="demo_group",
        agent_id="demo_agent",
        user_id="user123"
    )

    # Add memory
    memory.add("Test memory")

except Exception as e:
    print(f"Error: {e}")

Context Manager Usage

from memmachine_client import MemMachineClient

# Use client as context manager
with MemMachineClient(base_url="http://localhost:8080") as client:
    memory = client.memory(
        group_id="demo_group",
        agent_id="demo_agent",
        user_id="user123"
    )

    memory.add("This is a test memory")
    results = memory.search("test")
    print(f"Results: {results}")

# Client is automatically closed

Configuration

Environment Variables

• MEMORY_BACKEND_URL: Base URL for MemMachine server (default: http://localhost:8080)
• MEMORY_API_KEY: API key for authentication (optional for local development)

Client Configuration

client = MemMachineClient(
    api_key="your_api_key",  # Optional
    base_url="http://localhost:8080",
    timeout=30,  # Request timeout in seconds
    max_retries=3  # Maximum retries for failed requests
)

Running Examples

# Start MemMachine server first
python -m memmachine_server.server.app

# Run examples
python examples/memmachine_client_example.py

Contributing

Contributions are welcome! Please feel free to submit issues and pull requests.

License

This project is licensed under the same license as MemMachine.