# neo4j-contrib/mcp-neo4j [Official] [Health: Active]

**Category:** 🗄️ Databases  
**Repository:** https://github.com/neo4j-contrib/mcp-neo4j  
**GitHub Stars:** 989  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/neo4j-contrib-mcp-neo4j

## Description
Model Context Protocol with Neo4j (Run queries, Knowledge Graph Memory, Manaage Neo4j Aura Instances)

## Claude Desktop Quick Installation
No generated config for this server: it installs through its own setup steps.

The project's own setup steps (from https://github.com/neo4j-contrib/mcp-neo4j), copied verbatim. Run them in order; they are not an mcpServers config.

1. Start the server:

```sh
mcp-neo4j-cypher --transport http
```

2. Start the server:

```sh
mcp-neo4j-cypher --transport http --host 127.0.0.1 --port 8080 --path /api/mcp/
```

3. Start the server:

```sh
mcp-neo4j-cypher
```

## Documentation

## What neo4j-contrib/mcp-neo4j MCP server does

The neo4j-contrib/mcp-neo4j MCP server is a collection of Neo4j Labs MCP servers rather than one narrowly scoped database connector. It connects MCP clients such as Claude Desktop, Cursor, and Windsurf with Neo4j databases, personal graph memory, Neo4j Aura, and graph data modeling workflows.

The repository includes four server components:

- `mcp-neo4j-cypher` provides database schema inspection and executes generated read and write Cypher queries against a configured database. Schema inspection requires the APOC plugin to be installed and enabled on the Neo4j instance.
- `mcp-neo4j-memory` stores and retrieves entities and relationships in a personal knowledge graph backed by a local or remote Neo4j instance. The information can be reused across sessions, conversations, and clients.
- `mcp-neo4j-cloud-aura-api` manages Neo4j Aura instances, including listing, finding by name, creating, destroying, scaling, and enabling features.
- `mcp-neo4j-data-modeling` supports interactive graph data model creation, validation, visualization, and import or export through Arrows.app.

## How it works

An MCP client sends natural-language requests to the selected server component. The Cypher server translates requests into Cypher and runs them against the configured database. The memory server persists graph facts in Neo4j, while the Aura API server operates on cloud instances and the modeling server works with graph model representations.

All components support STDIO as the default transport, which is suitable for local integrations such as Claude Desktop. SSE is available for web-based deployments, and HTTP provides streamable communication for modern web deployments and microservices.

## Setup and configuration

The neo4j-contrib/mcp-neo4j MCP server can be run in HTTP mode by passing `--transport http`. The example configuration also supports `--host`, `--port`, and `--path` flags. Equivalent environment variables are `NEO4J_TRANSPORT`, `NEO4J_MCP_SERVER_HOST`, `NEO4J_MCP_SERVER_PORT`, and `NEO4J_MCP_SERVER_PATH`.

It does show component commands such as `mcp-neo4j-cypher --transport http`; the appropriate component must be selected for the intended task. Cloud deployment is supported through containerized HTTP deployments, with documentation covering AWS ECS Fargate and Azure Container Apps.

## Tools and capabilities

Use the Cypher component for graph exploration, schema-aware query generation, and database mutations. Use memory when facts about people, work, or other entities need to persist in a Neo4j-backed personal knowledge graph. Use the Aura API component for instance administration, including capacity changes and feature enablement. Use data modeling when a graph schema needs to be created, checked, visualized, or exchanged with Arrows.app.

## Limitations and notes

The neo4j-contrib/mcp-neo4j MCP server is part of the Neo4j Labs program. No service-level agreements or backward-compatibility and deprecation guarantees are provided.

The repository is distinct from Neo4j's separate official MCP server. HTTP deployment guidance is provided, but production security and credential details must be configured according to the deployment environment.

_Full upstream README: https://allmcps.com/mcp/neo4j-contrib-mcp-neo4j/readme_