MCP Memory Server

🚀 Knowledge Graph Memory Server

A basic in-memory implementation for knowledge graph persistence, leveraging local knowledge graphs. This enables Claude to retain information from user conversations.

⚠️ Important Note

This project is extracted from modelcontextprotocol/servers to create an independent implementation.

✨ Features

Entities

Entities are the core elements in the knowledge graph. Each entity has a unique identifier and a set of properties.

Structure:

• id: A unique identifier.
• type: The type (e.g., person, organization, location, etc.).
• properties: Attributes, including:
  • name: The name.
  • description: The description.
  • relations: A list of relationships.

Example:

{
  "id": "user_123",
  "type": "person",
  "properties": {
    "name": "张三",
    "age": 30,
    "job": "Software Engineer"
  }
}

Relationships

Relationships represent the connections between entities, with directionality and weight.

Structure:

• id: A unique identifier.
• from: The ID of the source entity.
• to: The ID of the target entity.
• type: The type (e.g., knows, works_at, lives_in).
• weight: The weight (default is 1).

Example:

{
  "id": "relation_456",
  "from": "user_123",
  "to": "company_789",
  "type": "works_at",
  "weight": 0.8
}

📚 Documentation

API

Create an Entity

async createEntity(properties) => Entity

Parameters:

• properties: The properties of the entity.

Returns:

• The newly created entity object.

Query an Entity

async getEntity(entityId) => Optional<Entity>

Parameters:

• entityId: The ID of the entity.

Returns:

• If the entity exists, return the entity; otherwise, return null.

Update an Entity

async updateEntity(entityId, properties) => Entity

Parameters:

• entityId: The ID of the entity.
• properties: The new properties.

Returns:

• The updated entity object.

Delete an Entity

async deleteEntity(entityId) => boolean

Parameters:

• entityId: The ID of the entity.

Returns:

• Return true if the deletion is successful; otherwise, return false.

📦 Installation

Use as an Independent Package

Install dependencies:

npm install mcp-memory-server

Run the command:

npx mcp-memory-server

Docker Deployment

Build the image:

docker build -t mcp/memory -f src/Dockerfile .

📄 License

This MCP server is released under the MIT License. This means you are free to use, modify, and distribute the software, subject to the terms and conditions of the MIT License. For more detailed information, please refer to the LICENSE file in the project repository.

💻 Usage Examples

System Prompt Example (for Claude's Custom Instructions)

Follow these steps for each interaction:

1. User Identification:
   - Assume you are interacting with the default_user.
   - If the default_user has not been determined, actively attempt to identify them.

2. Memory Retrieval:
   - Always start by saying "Remembering...", and retrieve all relevant information from the knowledge graph.
   - Refer to your knowledge graph as "memory".

3. Memory:
   - When communicating with the user, pay attention to any new information, including the following categories:
     a) Basic personality (age, gender, location, position, education level, etc.)
     b) Behavior (interests, habits, etc.)
     c) Preferences (communication style, preferred language, etc.)
     d) Goals (objectives, targets, aspirations, etc.)
     e) Relationships (personal and professional relationships, up to three levels)

4. Memory Update:
   - If any new information is collected during the interaction, update the memory as follows:
     a) Create entities for recurring organizations, people, and important events.
     b) Connect them to the current entities using relationships.
     c) Store facts as observations.