knowledge-mcp

What is knowledge-mcp?

knowledge-mcp is a prototype MCP server that serves knowledge from a Neo4J database, using dummy guidelines as example documents. It is intended for developers working with MCP-compatible clients like Cursor or Claude Desktop.

How to use knowledge-mcp?

Run Neo4J in a Docker container, run the included data loader to populate documents, build the MCP server binary, then configure it in an MCP client (e.g., Cursor) by providing the server path, OpenAI API key, and Neo4J credentials. The server communicates over stdio transport.

Key features of knowledge-mcp

• Serves knowledge from a Neo4J database
• Uses OpenAI API for processing questions
• Prototype – not ready for production
• Configured via stdio transport
• Works with Cursor, Claude Desktop, and other MCP clients
• Includes an example data loader

Use cases of knowledge-mcp

• Answer questions based on internal documents stored in Neo4J
• Prototype knowledge base retrieval for MCP clients
• Test Neo4J integration with the MCP ecosystem

FAQ from knowledge-mcp

What is the purpose of this server?

It demonstrates how knowledge can be extracted from a Neo4J database via the Model Context Protocol. The included documents are dummy guidelines.

Is knowledge-mcp ready for production?

No, it is a prototype and is not ready for production use.

What dependencies are required?

You need a running Neo4J instance (e.g., via Docker), an OpenAI API key, and you must build the .NET 9 server binary.

How do I configure knowledge-mcp in Cursor?

Add an MCP server in Cursor Settings with type stdio, command set to the built server binary path, and args containing your OpenAI key, Neo4J URI (neo4j://localhost:7687), username (neo4j), and password.

Can I use knowledge-mcp with other clients?

Yes, it works in any client that supports MCP, such as Claude Desktop.