mcpField
K
knowledge-mcp

knowledge-mcp

Local
github.com/simon-kMemory & Knowledge3Updated 79w ago
Repository

About knowledge-mcp

knowledge-mcp is a prototype MCP server that serves knowledge stored in Neo4J, using documents loaded into the database to answer questions. It is intended for MCP clients querying document-based knowledge; the README notes it is not production-ready.

Key features

  1. 01Serves knowledge from Neo4J
  2. 02Uses loaded documents to answer questions
  3. 03Prototype, not production-ready

Use cases

  1. 01Querying document knowledge through an MCP client
  2. 02Demonstrating knowledge extraction from Neo4J

MCP support

Exposes knowledge from Neo4J to MCP clients, which can use it to answer questions based on loaded documents.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
  "mcpServers": {
    "knowledge": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "\\"
      ]
    }
  }
}
Run commanddocker run -i --rm '\'

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
oci\

Tools

No tool list yet

Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.

Details

From the project README

knowledge-mcp

MCP Server that serves knowledge from a Neo4J Database.

The documents provided in this example are dummy guidelines. They are not real and serve the purpose to demonstrate that knowledge can be extracted from Neo4J.

This is a prototype and is not ready for production use!

Getting Started

Run Neo4J in a Docker Container

Run Neo4J Docker Container

docker run \
    --restart always \
    --publish=7474:7474 \
    --publish=7687:7687 \
    neo4j:latest

Login to Neo4J in the browser http://localhost:7474/ and set a new password for the user neo4j.

Run the Data Loader

Run the Data Load er project to save some documents in Neo4J.These documents will be used by the MCP server to answer questions.

Remember to change the source code so that it points to the Neo4J instance you are running and the Neo4J username and password is set according to the one you set in the previous step.

Build the MCP Server

Build the server to generate the binaries that are needed.

Configure the MCP Server in Cursor

Go to Cursor Settings and add a new MCP server.

Remember to set the YOUR_PATH, YOUR_OPENAI_API_KEYand YOUR_NEO4J_PASSWORD to the correct values.

{
  "mcpServers": {
    "KnowledgeBase": {
      "type": "stdio",
      "command": "YOUR_PATH/knowledge-mcp/src/KnowledgeMcp.Server/bin/Debug/net9.0/KnowledgeMcp.Server",
      "args": [
        "YOUR_OPEN_AI_API_KEY",
        "neo4j://localhost:7687",
        "neo4j",
        "YOUR_NEO4J_PASSWORD"
      ]
    }
  }
}

Notes

  • You can use the MCP server in any client that supports MCP. Fx the Claude Desktop.
  • A possible first improvement could be to use Neo4Js response format to get a citation to the document used to answer the question.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.