Neo4j Knowledge Graph

🚀 MCP Neo4j Memory Server

The MCP Neo4j Memory Server is a knowledge graph storage and retrieval tool based on the Neo4j graph database. It is designed with memory enhancement features for the Model Context Protocol (MCP). This server can handle entity creation, relationship establishment, and storage of observation information, and supports complex graph data queries.

🚀 Quick Start

The MCP Neo4j Memory Server is a powerful tool for storing and retrieving knowledge graphs. It offers a range of features to handle various tasks related to entity management and graph data queries.

✨ Features

• Knowledge Graph Storage: Efficiently stores entities and their relationships in the Neo4j graph database.
• Complex Query Capability: Utilizes the Cypher query language for efficient graph traversal and pattern matching.
• Fuzzy Search: Combines Fuse.js and Neo4j full - text indexing to provide flexible entity search functionality.
• Scalability: Supports cluster deployment to handle large - scale knowledge graphs.

📦 Installation

Install via npm:

npm install @jovanhsu/mcp-neo4j-memory-server

💻 Usage Examples

Basic Usage

Create entities and their relationships:

import { McpClient } from '@modelcontextprotocol/sdk/client/mcp.js';
import { StdioClientTransport } from '@modelcontextprotocol/sdk/client/stdio.js';

// Create client transport
const transport = new StdioClientTransport({
  command: 'npx',
  args: ['-y', '@jovanhsu/mcp-neo4j-memory-server'],
  env: {
    NEO4J_URI: 'bolt://localhost:7687',
    NEO4J_USER: 'neo4j',
    NEO4J_PASSWORD: 'password'
  }
});

const client = new McpClient();
await client.connect(transport);

// Create entities
const createResult = await client.callTool('create_entities', {
  entities: [
    {
      name: '李明',
      entityType: '人物',
      observations: ['喜欢编程', '使用TypeScript']
    },
    {
      name: 'Google',
      entityType: '公司',
      observations: ['总部位于美国', '业务范围包括搜索、云计算等']
    }
  ]
});

console.log('Create entities result:', createResult);

// Create a relationship
const relationResult = await client.callTool('create_relation', {
  sourceEntityName: '李明',
  targetEntityName: 'Google',
  relationType: '工作于'
});

console.log('Create relation result:', relationResult);

🔧 Technical Details

Data Model

The storage model of the knowledge graph in Neo4j is as follows:

(Entity:EntityType {name: "实体名称"}) Create entities of different types.
(Entity)-[:HAS_OBSERVATION]->(Observation {content: "观察内容}) Store observation information related to the entity.
(Entity1)-[:RELATION_TYPE]->(Entity2) Establish complex relationships between entities.

Fuzzy Search Implementation

Combines Neo4j's full - text indexing and Fuse.js for fuzzy matching:

• Uses Neo4j's full - text indexing for preliminary search.
• Provides additional fuzzy matching capabilities through Fuse.js.
• Search results include exact and partial matches, sorted by relevance.

🛠️ Development

Environment Configuration

# Clone the repository
git clone https://github.com/JovanHsu/mcp-neo4j-memory-server.git
cd mcp-neo4j-memory-server

# Install dependencies
pnpm install

# Build the project
pnpm build

# Development mode (using MCP Inspector)
pnpm dev

Testing

# Run tests
pnpm test

Why Choose Neo4j?

Compared to the original version using JSON file storage and the DuckDB version, Neo4j has the following advantages:

1. Native Graph Database: Neo4j is a database designed specifically for graph data, making it ideal for knowledge graph storage and querying.
2. High - Performance Queries: The Cypher query language enables efficient complex graph traversal and pattern matching.
3. Relationship - First: Neo4j treats relationships as first - class citizens, making relationship queries between entities more efficient.
4. Visualization Capability: Neo4j provides built - in visualization tools for easy debugging and understanding of the knowledge graph.
5. Scalability: Supports cluster deployment to handle large - scale knowledge graphs.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

📞 Contact

• GitHub: https://github.com/JovanHsu/mcp-neo4j-memory-server
• NPM: https://www.npmjs.com/package/@jovanhsu/mcp-neo4j-memory-server