neo4j-contrib/mcp-neo4j

What it solves

This project provides a set of Model Context Protocol (MCP) servers that allow Large Language Models (LLMs) to interact directly with Neo4j graph databases and cloud services. It bridges the gap between natural language interfaces (like Claude Desktop or Gemini CLI) and the technical operations required to query, manage, and model data in Neo4j.

How it works

It implements the MCP standard to create a communication layer between an AI client and Neo4j. The repository contains several specialized servers:

• mcp-neo4j-cypher: Converts natural language into Cypher queries to read and write data.
• mcp-neo4j-memory: Uses Neo4j as a persistent knowledge graph for storing and retrieving entities and relationships across different AI sessions.
• mcp-neo4j-cloud-aura-api: Provides an interface to manage Neo4j Aura cloud instances (creating, scaling, and destroying instances).
• mcp-neo4j-data-modeling: Enables the creation, validation, and visualization of graph data models.

These servers support multiple transport modes, including STDIO for local use and SSE/HTTP for cloud deployments on platforms like AWS ECS Fargate or Azure Container Apps.

Who it’s for

Developers and data engineers who want to integrate Neo4j graph databases into their AI workflows, allowing LLMs to perform database operations, maintain long-term memory, or manage cloud infrastructure via chat.

Highlights

• Multi-server ecosystem: Specialized tools for querying, memory, cloud management, and modeling.
• mcp-neo4j-memory: Enables persistent, cross-session knowledge graph memory for AI agents.
• Cloud-ready: Containerized servers with dedicated HTTP transport for scalable production deployments.
• Standardized Protocol: Built on the Model Context Protocol for compatibility with various MCP clients like Cursor, Windsurf, and Claude Desktop.