Skip to content

Latest commit

 

History

History
239 lines (187 loc) · 6.41 KB

README.md

File metadata and controls

239 lines (187 loc) · 6.41 KB

CrewAI Integration with MemMachine

This directory contains tools for integrating MemMachine with CrewAI agents.

Overview

MemMachine provides memory tools that can be integrated into CrewAI agents to enable persistent memory capabilities. This allows agents to remember past interactions, user preferences, and context across multiple sessions.

Installation

# Install CrewAI and crewai-tools
pip install crewai crewai-tools

# Install MemMachine client
pip install memmachine-client

Configuration

The integration can be configured via environment variables:

Variable Description Default
MEMORY_BACKEND_URL URL of the MemMachine backend service http://localhost:8080
CREWAI_ORG_ID Organization identifier crewai_org
CREWAI_PROJECT_ID Project identifier crewai_project
CREWAI_GROUP_ID Group identifier (optional) None
CREWAI_AGENT_ID Agent identifier (optional) None
CREWAI_USER_ID User identifier (optional) None
CREWAI_SESSION_ID Session identifier (optional) None

Quick Start

1. Basic Usage

from crewai import Agent, Crew, Task
from integrations.crewai.tool import create_memmachine_tools

# Create MemMachine tools
memmachine_tools = create_memmachine_tools(
    base_url="http://localhost:8080",
    org_id="my_org",
    project_id="my_project",
    user_id="user123",
)

# Create an agent with memory tools
researcher = Agent(
    role="Researcher",
    goal="Research topics and remember important information",
    backstory="You are a helpful research assistant with memory capabilities.",
    tools=memmachine_tools,
    verbose=True,
)

# Create a task
task = Task(
    description="Research AI trends and remember key findings",
    agent=researcher,
)

# Create and run the crew
crew = Crew(
    agents=[researcher],
    tasks=[task],
)

result = crew.kickoff(inputs={"topic": "Latest AI trends"})

2. Advanced Usage with Multiple Agents

from crewai import Agent, Crew, Task
from integrations.crewai.tool import create_memmachine_tools

# Create shared memory tools
memmachine_tools = create_memmachine_tools(
    base_url="http://localhost:8080",
    org_id="my_org",
    project_id="my_project",
    group_id="research_team",
)

# Create multiple agents with memory
researcher = Agent(
    role="Researcher",
    goal="Research and store findings in memory",
    backstory="You research topics and remember important information.",
    tools=memmachine_tools,
    verbose=True,
)

writer = Agent(
    role="Writer",
    goal="Write content based on research and recall past work",
    backstory="You write articles and can recall previous research.",
    tools=memmachine_tools,
    verbose=True,
)

# Create tasks
research_task = Task(
    description="Research AI trends and store findings in memory",
    agent=researcher,
)

writing_task = Task(
    description="Search memory for previous research and write an article",
    agent=writer,
)

# Create and run the crew
crew = Crew(
    agents=[researcher, writer],
    tasks=[research_task, writing_task],
)

result = crew.kickoff(inputs={"topic": "AI in healthcare"})

3. Using MemMachineTools Directly

from integrations.crewai.tool import MemMachineTools

# Initialize tools
tools = MemMachineTools(
    base_url="http://localhost:8080",
    org_id="my_org",
    project_id="my_project",
    user_id="user123",
)

# Add memory
from memmachine_server.common.api import EpisodeType

result = tools.add_memory(
    content="User prefers Python over JavaScript",
    role="user",
    episode_type=EpisodeType.MESSAGE,  # Optional: defaults to EpisodeType.MESSAGE
)
print(result)

# Search memory
results = tools.search_memory(
    query="What programming languages does the user prefer?",
    limit=5,
)
print(results["summary"])

Tool Descriptions

Add Memory Tool

Stores important information, facts, preferences, or conversation context in MemMachine memory. Use this automatically whenever the user shares new information that should be remembered.

Parameters:

  • content: The content to store in memory (required)
  • role: Message role - "user", "assistant", or "system" (default: "user")

Search Memory Tool

Retrieves relevant context, past conversations, or user preferences from MemMachine memory. Use this when you need to recall information from previous interactions.

Parameters:

  • query: Search query string describing what you're looking for (required)
  • limit: Maximum number of results to return (default: 5)

Example: Research Agent with Memory

from crewai import Agent, Crew, Task
from integrations.crewai.tool import create_memmachine_tools
import os

# Get configuration from environment
base_url = os.getenv("MEMORY_BACKEND_URL", "http://localhost:8080")
org_id = os.getenv("CREWAI_ORG_ID", "research_org")
project_id = os.getenv("CREWAI_PROJECT_ID", "research_project")

# Create memory tools
memmachine_tools = create_memmachine_tools(
    base_url=base_url,
    org_id=org_id,
    project_id=project_id,
    user_id="researcher_001",
)

# Create agent with memory
researcher = Agent(
    role="Research Assistant",
    goal="Research topics thoroughly and remember key findings",
    backstory="""You are an expert researcher with excellent memory.
    You always store important findings in memory so you can recall them later.
    When researching, you search memory first to see if you've researched this topic before.""",
    tools=memmachine_tools,
    verbose=True,
    allow_delegation=False,
)

# Create task
task = Task(
    description="""Research the topic: {topic}

    1. First, search memory to see if you've researched this topic before
    2. Research the topic thoroughly
    3. Store key findings in memory for future reference
    4. Provide a comprehensive summary""",
    agent=researcher,
)

# Create and run crew
crew = Crew(
    agents=[researcher],
    tasks=[task],
)

result = crew.kickoff(inputs={"topic": "Quantum Computing"})
print(result)

Requirements

Notes

  • Memory tools are shared across all agents in a crew by default
  • Each agent can have its own agent_id in metadata for tracking
  • Use user_id to scope memories to specific users
  • Use session_id to scope memories to specific sessions
  • Memories persist across crew runs, enabling long-term context