M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
N

Neo4j Data Modeling MCP Server

Community
Create, validate, and visualize Neo4j graph data models
How to install ↓ Compare GitHub source repository ↗
Category
Database #37 of 58
Stars
★ 988 Very popular
Transport
stdio (local process)
Runtime
Python · Docker
Credentials
No credential needed
License
MIT
Last commit
51FMRS · D

A Neo4j Labs graph data modeling MCP server suited to designing and validating models before committing them, but experimental with no official product support or compatibility guarantees.

Strongest · Maintenance 13/20 Weakest · Reliability 8/20

Reliability
8/20
Security and permissions
10/20
Maintenance
13/20
Documentation
9/20
Setup experience
11/20
Why each score
Reliability 8/20
This static review is limited to the supplied manifest and repository-level README; it contains no source code, unit tests, or CI workflows for servers/mcp-neo4j-data-modeling. Per the static calibration, without verifiable execution evidence (real CI plus committed tests covering key paths) reliability cannot exceed 12. The manifest declares stdio transport and one optional environment variable (NEO4J_NAMESPACE), the package name matches the description, and no contradictions or hidden external dependencies (database, network) are visible, so a clean startup and handshake look plausible. However, no tool list, input/output schema, error handling, or dependency resolution evidence exists, so 'declared tools match real behavior' cannot be verified; edge cases and failure diagnosis are unproven. Scored at 8, the 'happy path plausible but tests/edge cases/error handling thin' anchor.
Security and permissions 10/20
No red lines triggered: no malware, credential theft, covert exfiltration, unconfirmed destructive delete/pay/deploy/remote-command defaults, fabricated maintainer identity, or real secrets in install examples. The only environment variable, NEO4J_NAMESPACE, is explicitly optional and non-secret, so the permission surface is small; a data-modeling server does not inherently need Neo4j credentials, which favors least privilege. Deductions are from missing evidence: the files do not show whether model import/export (e.g., Arrows.app) touches external networks, writes to disk or a remote Neo4j, whether dangerous operations require confirmation, or how data boundaries are disclosed. Scored at 10, the 'main risks visible but permissions/confirmation/scoping incomplete' anchor, lowered for uncertainty.
Maintenance 13/20
The repository belongs to the neo4j-contrib GitHub organization as a Neo4j Labs project, explicitly developed and maintained by the Neo4j Field GenAI team with community contributions welcomed, under an MIT license, not archived, with 980 stars and 25 open issues indicating ongoing attention. The manifest pins a concrete semver version 0.8.2 and a PyPI identifier, giving a clear release path. Deductions: Labs status explicitly disclaims any SLA and makes no guarantees about backward compatibility or deprecation, so governance/stability commitments are incomplete; there is no evidence of dependency-update policy, security-response channel, or an independent release cadence for this sub-server. Scored at 13, the 'active but governance/versioning gaps' anchor.
Documentation 9/20
The repository README offers a four-server overview, transport-mode explanations (STDIO/SSE/HTTP), and HTTP configuration examples and links a cloud deployment guide plus blog posts, giving some layered structure and honest disclosure of the no-SLA Labs limitation. Deductions are significant: the top-level README describes sibling servers, and its example commands (e.g., mcp-neo4j-cypher --transport http) cannot be attributed to the data-modeling sub-server. For this server there is only one line ('create, validate, and visualize Neo4j graph data models; import/export from Arrows.app'), with no tool parameters, return values, auth requirements, limits, cost, or troubleshooting content. Scored at 9, the 'usable but hidden assumptions and troubleshooting gaps' anchor.
Setup experience 11/20
The install path is fairly clear: PyPI package mcp-neo4j-data-modeling 0.8.2 over stdio, with a single optional environment variable that can be omitted for defaults, implying few steps and low complexity; the README also names mainstream clients such as Claude Desktop, VS Code, and Cursor. But the manifest contains no client configuration example (e.g., a claude_desktop_config snippet), and the README's HTTP/environment-variable examples target other servers, so nothing copy-pasteable for this server is provided; platform compatibility, Python version, and containerization are also undocumented. Per static calibration, setup cannot exceed 15 without verifiable execution evidence; evidence here is thin and partially misleading toward sibling servers, so 11 is awarded.

Static review · not runListed 2026-09-10

Read the FMRS scoring method →

Fit and risk

What it can accessConnects to a database

Best for

  • Neo4j users who want to design and validate graph models before committing them
  • Developers doing graph data modeling from MCP-compatible clients

Not for

  • Teams requiring official Neo4j product support or SLAs
  • Production environments requiring backwards-compatibility guarantees

Required permissions

  • Runs locally over stdio using local machine resources

Risks and side effects

  • Neo4j Labs project with frequently updated experimental features and no guarantees around backwards compatibility or deprecation
  • Not supported by the Neo4j product team
  • Importing/exporting models from Arrows.app may send model data to an external service

Setup

Before you start

Runtime：Python · Docker

Other optional settings (1)
NEO4J_NAMESPACE optional Optional; sets the namespace used for the MCP server's tool names. No credential needed.

Install mcp-neo4j-data-modeling (version 0.8.2) from PyPI and configure it in an MCP-compatible client using stdio transport. Optionally set NEO4J_NAMESPACE to control tool-name namespacing.

Check that it works

After install, check that the mcp-neo4j-data-modeling server's data modeling tools appear in your MCP client's tool list; alternatively ask the assistant to create and visualize a simple graph model — returning the model proves the connection works.

Troubleshooting

  1. Confirm version 0.8.2 of mcp-neo4j-data-modeling is installed
  2. Confirm the MCP client is configured for stdio transport
  3. Check whether NEO4J_NAMESPACE is set as expected for tool naming

Things to try

Once connected, you can ask your AI assistant things like:

  • Create a graph data model for users, orders and products in Neo4j
  • Validate my existing graph data model against Neo4j best practices
  • Export the current graph data model for import into Arrows.app
  • Visualize the nodes and relationships in this data model

Use cases

Interactively design Neo4j graph data models
Validate graph data models before committing them
Visualize graph data model structures
Import and export data models from Arrows.app

Overview

This server is one of four MCP servers in the neo4j-contrib/mcp-neo4j repository, developed by the Neo4j Field GenAI team as part of the Neo4j Labs program and not supported by the Neo4j product team. It provides interactive graph data modeling and visualization: create, validate, and visualize Neo4j graph data models, with model import/export from Arrows.app. It runs as the PyPI package mcp-neo4j-data-modeling (version 0.8.2) over stdio transport, with an optional NEO4J_NAMESPACE environment variable for tool-name namespacing.

Similar servers

Compare with other MCP servers

More servers from this repository

Neo4j Aura Manager MCP ServerNeo4j Cypher MCP ServerNeo4j Knowledge Graph Memory Server

Source revision 1c7ad0793b5f Data synced 2026-10-19 Read the FMRS scoring method