mcp/neo4j-data-modeling repository overview

⁠Neo4j Data Modeling MCP Server

MCP server that assists in creating, validating and visualizing graph data models.

What is an MCP Server?⁠

⁠MCP Info

⁠Image Building Info

Attribute	Details
Dockerfile	https://github.com/neo4j-contrib/mcp-neo4j/blob/dbc01ba78f171851f2d57dcd125b028c29912fd1/servers/mcp-neo4j-data-modeling/Dockerfile⁠
Commit	dbc01ba78f171851f2d57dcd125b028c29912fd1
Docker Image built by	Docker Inc.
Docker Scout Health Score
Verify Signature	COSIGN_REPOSITORY=mcp/signatures cosign verify mcp/neo4j-data-modeling --key https://raw.githubusercontent.com/docker/keyring/refs/heads/main/public/mcp/latest.pub
Licence	MIT License

⁠Available Tools (16)

⁠Tools Details

⁠Tool: export_to_arrows_json

Export the data model to the Arrows web application format. Returns a JSON string. This should be presented to the user as an artifact if possible. Accepts either a DataModel object or a JSON string of the DataModel object.

This tool is read-only. It does not modify its environment.

⁠Tool: export_to_neo4j_graphrag_pkg_schema

Export a data model to a Neo4j Graphrag Python Package schema. Returns a dictionary containing the Neo4j Graphrag Python Package schema. Accepts either a DataModel object or a JSON string of the DataModel object.

This tool is read-only. It does not modify its environment.

⁠Tool: export_to_owl_turtle

Export a data model to an OWL Turtle string. This process is lossy since OWL does not support properties on relationships. Returns a string representation of the data model in OWL Turtle format. Accepts either a DataModel object or a JSON string of the DataModel object.

This tool is read-only. It does not modify its environment.

⁠Tool: export_to_pydantic_models

Export a data model to Pydantic models. Returns a string representation of the Pydantic models for the data model as a Python file. Accepts either a DataModel object or a JSON string of the DataModel object.

This tool is read-only. It does not modify its environment.

⁠Tool: get_constraints_cypher_queries

Get the Cypher queries to create constraints on the data model. This creates range indexes on the key properties of the nodes and relationships and enforces uniqueness and existence of the key properties. Accepts either a DataModel object or a JSON string of the DataModel object.

This tool is read-only. It does not modify its environment.

⁠Tool: get_example_data_model

Get an example graph data model from the available templates. Returns a DataModel object and the Mermaid visualization configuration for the example graph data model.

This tool is read-only. It does not modify its environment.

⁠Tool: get_mermaid_config_str

Get the Mermaid configuration string for the data model. This may be visualized in Claude Desktop and other applications with Mermaid support. Accepts either a DataModel object or a JSON string of the DataModel object.

This tool is read-only. It does not modify its environment.

⁠Tool: get_node_cypher_ingest_query

Get the Cypher query to ingest a list of Node records into a Neo4j database. This should be used to ingest data into a Neo4j database. This is a parameterized Cypher query that takes a list of records as input to the $records parameter.

This tool is read-only. It does not modify its environment.

⁠Tool: get_relationship_cypher_ingest_query

Get the Cypher query to ingest a list of Relationship records into a Neo4j database. This should be used to ingest data into a Neo4j database. This is a parameterized Cypher query that takes a list of records as input to the $records parameter. The records must contain the Relationship properties, if any, as well as the sourceId and targetId properties of the start and end nodes respectively.

This tool is read-only. It does not modify its environment.

⁠Tool: list_example_data_models

List all available example data models with descriptions. Returns a dictionary with example names and their descriptions.

⁠Tool: load_from_arrows_json

Load a data model from the Arrows web application format. Returns a data model as a JSON string.

This tool is read-only. It does not modify its environment.

⁠Tool: load_from_neo4j_graphrag_pkg_schema

Load a data model from a Neo4j Graphrag Python Package schema. Returns a DataModel object. Accepts a Neo4j Graphrag Python Package schema dictionary.

This tool is read-only. It does not modify its environment.

⁠Tool: load_from_owl_turtle

Load a data model from an OWL Turtle string. This process is lossy and some components of the ontology may be lost in the data model schema. Returns a DataModel object.

This tool is read-only. It does not modify its environment.

⁠Tool: validate_data_model

Validate the entire data model. Returns True if the data model is valid, otherwise raises a ValueError. If return_validated is True, returns the validated data model. Accepts either a DataModel object or a JSON string of the DataModel object.

This tool is read-only. It does not modify its environment.

⁠Tool: validate_node

Validate a single node. Returns True if the node is valid, otherwise raises a ValueError. If return_validated is True, returns the validated node. Accepts either a Node object or a JSON string of the Node object.

This tool is read-only. It does not modify its environment.

⁠Tool: validate_relationship

Validate a single relationship. Returns True if the relationship is valid, otherwise raises a ValueError. If return_validated is True, returns the validated relationship. Accepts either a Relationship object or a JSON string of the Relationship object.

This tool is read-only. It does not modify its environment.

⁠Use this MCP Server

{
  "mcpServers": {
    "neo4j-data-modeling": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "NEO4J_TRANSPORT",
        "-e",
        "NEO4J_MCP_SERVER_HOST",
        "-e",
        "NEO4J_MCP_SERVER_PORT",
        "-e",
        "NEO4J_MCP_SERVER_PATH",
        "-e",
        "NEO4J_MCP_SERVER_ALLOW_ORIGINS",
        "-e",
        "NEO4J_MCP_SERVER_ALLOWED_HOSTS",
        "mcp/neo4j-data-modeling"
      ],
      "env": {
        "NEO4J_TRANSPORT": "http",
        "NEO4J_MCP_SERVER_HOST": "0.0.0.0",
        "NEO4J_MCP_SERVER_PORT": "8000",
        "NEO4J_MCP_SERVER_PATH": "/mcp/",
        "NEO4J_MCP_SERVER_ALLOW_ORIGINS": "example.com,www.example.com",
        "NEO4J_MCP_SERVER_ALLOWED_HOSTS": "localhost,127.0.0.1"
      }
    }
  }
}

Why is it safer to run MCP Servers with Docker?⁠