Version: 0.2.0

mem0 MCP Server bridges MCP Host applications and the mem0 cloud service, enabling structured project memory management and semantic search for project-related information.

Release Notes

v0.2.0

• Switched from SSE-based to stdio-based invocation for better compatibility with MCP Hosts
• Added support for pipx-based installation and execution
• Simplified deployment via pyproject.toml script entrypoint

Features

• Project memory storage and retrieval
• Semantic search for project information
• Structured project management data handling
• Fully tested stdio-based MCP Server tools
• Flexible logging: stderr by default, file output via --logfile
• Smart CLI invocation via pipx-compatible interface

MCP Host Configuration

Example configuration for MCP Host:

"mem0": {
  "command": "pipx",
  "args": ["run", "mem0-mcp-for-pm"],
  "env": {
    "MEM0_API_KEY": "{apikey}"
  }
}

Tools

• add_project_memory
• get_all_project_memories
• search_project_memories
• update_project_memory
• delete_project_memory
• delete_all_project_memories

All tools are available via stdio-based MCP protocol.

Logging

• Default: stderr
• Optional: --logfile /path/to/logfile.log

License

See LICENSE file.

Technical details

The uniqueness of this forked is the structured format between MCP Host and mem0 is expected in coding format like Javascript object. Make sure you set the custom instruction to be able to handle better.

Custom instruction

In order to make mem0 working as fitting to project management purpose, this forked has the following instruction for AI.

For mem0

• Check the source code.

For MCP Host

• The following is just sample, find the best by yourself !!

mem0 Guide for Effective Project Memory (Enhanced)

This guide outlines strategies and templates for effectively managing project information using mem0. The aim is to improve searchability and reusability of project data through structured templates and metadata management.

Information Structure and Templates

mem0 can effectively manage the following types of information. Using structured templates improves searchability and reusability. Note that the templates provided are examples and should be adapted to fit specific project needs.

1. Project Status Management

Template:

// [PROJECT: project-name] [TIMESTAMP: yyyy-MM-ddTHH:mm:ss+09:00] [TYPE: Project Status]
const projectStatus = {
  overview: {
    name: "Project Name",      // Required
    purpose: "Project Purpose", // Required
    version: "1.2.0",          // Optional
    phase: "development"       // Optional
  },
  progress: {
    completionLevel: 0.65,    // Completion rate (value between 0 and 1)
    milestones: [
      { name: "Planning Phase", status: "completed", date: "2025-02-15" },
      { name: "Development Phase", status: "in-progress", progress: 0.70 }
    ]
  },
  currentFocus: ["Implementing Feature X", "Optimizing Component Y"],
  risks: ["Concerns about API stability", "Resource shortage"]
};

2. Task Management

Template:

// [PROJECT: project-name] [TIMESTAMP: yyyy-MM-ddTHH:mm:ss+09:00] [TYPE: Task Management]
const taskManagement = {
  highPriority: [
    {
      description: "Implement Feature X",     // Required
      status: "in-progress",                 // Required
      deadline: "2025-03-15",                // Optional
      assignee: "Team A",                    // Optional
      dependencies: "Component Y"            // Optional
    }
  ],
  mediumPriority: [],
  completedTasks: [
    {
      description: "Setup Development Environment",
      status: "completed"
    }
  ]
};

3. Meeting Summary

Template:

// [PROJECT: project-name] [TIMESTAMP: yyyy-MM-ddTHH:mm:ss+09:00] [TYPE: Meeting Summary]
const meetingMinutes = {
  title: "Weekly Progress Meeting",
  date: "2025-03-23",
  attendees: [
    { department: "Development", members: ["Sato", "Suzuki"] },
    { department: "Design", members: ["Tanaka"] }
  ],
  topics: ["Progress Report", "Risk Management", "Next Week's Plan"],
  decisions: [
    "Approve additional resource allocation",
    "Delay release date by one week"
  ],
  actionItems: [
    { description: "Procedure for adding resources", assignee: "Sato", dueDate: "2025-03-25" },
    { description: "Revise test plan", assignee: "Suzuki", dueDate: "2025-03-24" }
  ]
};

Effective Information Management Techniques

1. Context Management (run_id)

Using mem0's run_id parameter, you can logically group related information. This helps maintain specific conversation flows or project contexts.

Recommended Format:

project:project-name:category:subcategory

Usage Example:

// Managing information related to a specific feature
add_project_memory(
  "// [PROJECT: Member System] [TYPE: Technical Specification]
const authSpec = {...};",
  run_id="project:member-system:feature:authentication",
  metadata={"type": "specification"}
);

// Adding a task for the same feature
add_project_memory(
  "// [PROJECT: Member System] [TYPE: Task Management]
const authTasks = {...};",
  run_id="project:member-system:feature:authentication",
  metadata={"type": "task"}
);

// Searching for related information
search_project_memories("authentication", {
  "run_id": "project:member-system:feature:authentication"
});

2. Effective Use of Metadata

Using metadata can enhance the searchability of information. We recommend using the following schema:

{
  "type": "meeting|task|decision|status|risk", // Type of information
  "priority": "high|medium|low",               // Priority
  "tags": ["frontend", "backend", "design"],   // Related tags
  "status": "pending|in-progress|completed"    // Status
}

Usage Example:

// Registering a high-priority task
add_project_memory(
  "// [PROJECT: Member System] [TYPE: Task Management]
const task = {...};",
  metadata={
    "type": "task",
    "priority": "high",
    "tags": ["frontend", "authentication"]
  }
);

// Searching for tasks with a specific tag
search_project_memories("task", {
  "metadata": {
    "tags": ["frontend"]
  }
});

3. Information Lifecycle Management

Using the immutable and expiration_date parameters, you can manage the lifecycle of information.

Usage Example:

// Recording an immutable decision
add_project_memory(
  "// [PROJECT: Member System] [TYPE: Decision Record]
const decision = {...};",
  immutable=True,  // Set as immutable
  metadata={"type": "decision"}
);

// Information with an expiration date
add_project_memory(
  "// [PROJECT: Member System] [TYPE: Meeting Summary]
const meeting = {...};",
  expiration_date="2025-06-30",  // Expires on this date
  metadata={"type": "meeting"}
);

Practical Usage Patterns

1. Sprint Management Example

// Registering the sprint plan at the start
add_project_memory(
  "// [PROJECT: Member System] [TIMESTAMP: 2025-05-01T10:00:00+09:00] [TYPE: Project Status]
" +
  "const sprintPlan = {
" +
  "  sprint: "Sprint-2025-05",
" +
  "  duration: "2 weeks",
" +
  "  goals: ["Implement authentication feature", "Improve UI"],
" +
  "  tasks: [
" +
  "    { description: "Implement login screen", assignee: "Tanaka", estimate: "3 days" },
" +
  "    { description: "API integration", assignee: "Sato", estimate: "2 days" }
" +
  "  ]
" +
  "};",
  run_id="project:member-system:sprint:2025-05",
  metadata={"type": "status", "tags": ["sprint-planning"]}
);

// Mid-sprint progress report
add_project_memory(
  "// [PROJECT: Member System] [TIMESTAMP: 2025-05-08T15:00:00+09:00] [TYPE: Project Status]
" +
  "const progress = {
" +
  "  sprint: "Sprint-2025-05",
" +
  "  completionLevel: 0.4,
" +
  "  status: [
" +
  "    { task: "Implement login screen", progress: 0.7, status: "in-progress" },
" +
  "    { task: "API integration", progress: 0.2, status: "in-progress" }
" +
  "  ],
" +
  "  blockers: ["Change in API response specification"]
" +
  "};",
  run_id="project:member-system:sprint:2025-05",
  metadata={"type": "status", "tags": ["sprint-progress"]}
);

2. Risk Management Example

// Registering a risk
add_project_memory(
  "// [PROJECT: Member System] [TIMESTAMP: 2025-05-03T11:00:00+09:00] [TYPE: Risk Assessment]
" +
  "const risk = {
" +
  "  description: "Concerns about external API stability",
" +
  "  impact: "High",
" +
  "  probability: "Medium",
" +
  "  mitigation: "Implement fallback mechanism",
" +
  "  owner: "Development Lead"
" +
  "};",
  run_id="project:member-system:risk:api-stability",
  metadata={"type": "risk", "priority": "high"}
);

// Updating the risk status
add_project_memory(
  "// [PROJECT: Member System] [TIMESTAMP: 2025-05-10T16:30:00+09:00] [TYPE: Risk Assessment]
" +
  "const riskUpdate = {
" +
  "  description: "Concerns about external API stability",
" +
  "  status: "Resolved",
" +
  "  resolution: "Fallback mechanism implementation completed"
" +
  "};",
  run_id="project:member-system:risk:api-stability",
  metadata={"type": "risk", "priority": "medium"}
);

Important Points

• Standard Metadata: Always include the project name and timestamp.
• Data Format: Use structured data (JavaScript objects, JSON, YAML).
• Context Management: Use run_id hierarchically to maintain information relevance.
• Search Efficiency: Consistent metadata and structure improve search efficiency.

4. Implementation Strategy

To implement the above improvements, we recommend the following steps:

1. Enhance the add_project_memory Method:
2. Update documentation strings: Improve usage examples and parameter descriptions.
3. Error handling: Provide more detailed error information.

4. Response format: Explicitly state the parameters used.

5. Update Custom Instructions:

6. Enrich template examples.
7. Clarify recommended usage of run_id (introduce hierarchical structure).
8. Standardize metadata schema.
9. Provide practical usage examples.

These improvements will enhance the usability and efficiency of information management while maintaining compatibility with existing APIs.

5. Summary

The proposed improvements provide value in the following ways while maintaining compatibility with existing mem0 MCP server functions: