MCP Neo4j is a free MCP server tool developed by Neo4j Contrib for the mcp platform. Built with Python, it extends AI client capabilities by connecting to Neo4j graph databases. Distributed via PyPI packages and Docker images, it enables interaction with graph data through various callable tools. This program allows developers, data scientists, and application builders to integrate Neo4j as a knowledge source or reasoning backend in multi-agent systems, supporting diverse data operations and graph analysis.

What does MCP Neo4j connect to?

The MCP Neo4j tool connects directly to a Neo4j graph database. As an MCP server, it acts as an intermediary, allowing various AI clients to interact with graph data. This connection enables operations ranging from schema retrieval and Cypher query execution to detailed graph data science algorithms. Its design supports integration into environments where a Neo4j instance serves as a central data store for complex relationships and entities, making graph data accessible for AI-driven applications.

Core features and tools

The neo4j-contrib/mcp-neo4j project includes several MCP servers, each exposing specific callable tools. mcp-neo4j-cypher offers get_neo4j_schema, read_neo4j_cypher, and write_neo4j_cypher. mcp-neo4j-memory provides read_graph, search_nodes, find_nodes, create_entities, delete_entities, create_relations, delete_relations, add_observations, and delete_observations. mcp-neo4j-data-modeling includes validate_node. MCP-Neo4j-GDS contains many Graph Data Science algorithms, such as Article Rank, Betweenness Centrality, PageRank, and HITS.

Who is it for?

MCP Neo4j serves developers prototyping conversational Neo4j queries and data scientists exploring graph data without deep Cypher expertise. Platform teams can deploy structured, auditable Neo4j tooling, while application builders use Neo4j as a knowledge source in multi-agent systems. It supports various AI clients including Claude Desktop, VS Code, and JetBrains AI Assistant. Skip this if your project does not involve interaction with a Neo4j graph database.

How to get started

The server authenticates using an API key passed as an environment variable or per-request authentication in HTTP mode. For STDIO mode, credentials are provided via NEO4J_USERNAME and NEO4J_PASSWORD environment variables.

Use cases

• Text2Cypher workflows, enabling natural language interaction to execute read and write Cypher queries against a Neo4j database.
• Graph Data Modeling, allowing users to define nodes, relationships, and properties to design and visualize graph database schemas.
• Knowledge Graph Memory, storing and retrieving entities and relationships in a Neo4j instance for persistent memory across sessions.

MCP Neo4j connects AI clients to Neo4j graph databases, offering tools for Cypher interaction, data modeling, and graph data science. It supports multiple transport protocols and authentication methods, catering to diverse development and analytical needs. Its capabilities extend to natural language querying and persistent knowledge storage.