MCP Qdrant Memory

🚀 MCP Memory Server with Qdrant Persistence

This MCP server offers a semantic search feature based on a knowledge graph, powered by the Qdrant vector database.

✨ Features

• Graph-based knowledge representation, supporting entities and their relationships.
• File persistence (memory.json).
• Semantic search using the Qdrant vector database.
• Semantic similarity calculation with OpenAI's embedding models.
• Support for the HTTPS protocol and reverse proxy compatibility.
• Easy deployment in a Docker environment.

📦 Installation

Local Installation

1. Install dependencies:

npm install

2. Build the server:

npm run build

Docker Deployment

1. Build the Docker image:

docker build -t mcp-qdrant-memory .

2. Run the container with environment variables:

docker run -d \
  -e OPENAI_API_KEY=your-openai-api-key \
  -e QDRANT_URL=https://your-qdrant-server \
  -e QDRANT_API_KEY=your-qdrant-api-key \
  -e QDRANT_COLLECTION_NAME=your-collection-name \
  --name mcp-memory-server \
  mcp-qdrant-memory

💻 Usage Examples

Basic Usage

// Create two entities
await createEntity({ id: "1", name: "Alice", age: 30 });
await createEntity({ id: "2", name: "Bob", age: 25 });

// Query all entities
const entities = await getEntities();
console.log("All entities:", entities);

// Update an entity
await updateEntity("1", { age: 31 });
const alice = await getEntity("1");
console.log("Alice's updated age:", alice.age); // Output: 31

// Delete an entity
await deleteEntity("2");
const remainingEntities = await getEntities();
console.log("Remaining entities:", remainingEntities); // Output: [ { id: "1", name: "Alice", age: 31 } ]

Tool Interfaces

• createEntity(entityData: object): Promise
  Create and store entity data.
• getEntities(query?: string): Promise<Entity[]>
  Get a list of entities based on an optional query parameter. Returns all entities by default.
• updateEntity(entityId, entityData: object): Promise
  Update the attribute data of a specified entity.
• deleteEntity(entityId: string): Promise
  Delete a specific entity and its related relationships.

🔧 Technical Details

• Data Persistence: All entities and their relationships are stored in the memory.json file to ensure data integrity after service restarts.
• Search Mechanism: Utilize the Qdrant vector database for efficient semantic search, supporting text similarity matching and content retrieval.
• Error Handling: Incorporate a comprehensive transaction rollback mechanism to ensure data consistency during network fluctuations or service exceptions.

📚 Documentation

For a deeper understanding of the system's internal mechanisms or for secondary development, refer to the following resources:

1. Official Documentation
   MCP Memory Server Docs
2. Technical Blog Series
   • Best practices for knowledge graph construction.
   • Applications of Qdrant in semantic search.
   • High-availability architecture design.
3. Online Tutorials
   How to Build a Knowledge Retrieval System with the MCP Memory Server

📄 License

This project is licensed under the MIT Open Source License, allowing for free use, modification, and distribution.

⚠️ Security Considerations

• HTTPS Support: Implement secure communication via SSL/TLS certificates. HTTPS is enabled by default to ensure the security of data transmission.
• Authentication and Authorization: Authenticate using Qdrant's API keys to prevent unauthorized access and malicious attacks.
• Access Control: Set up an IP whitelist to restrict connections to the Qdrant service to specific sources, enhancing network security.

🤝 Contribution Guidelines

Contributions from the community are welcome to improve this project:

1. Clone the Repository

   git clone https://github.com/yourusername/mcp-memory-server.git
   

2. Report Issues and Suggestions Submit issues or feature suggestions in the GitHub Issues section.