Talk to Your Neo4j Database Like a Human

Stop context-switching between your AI assistant and the Neo4j browser. The Neo4j MCP server collection brings graph database operations directly into Claude Desktop, VS Code, and other MCP-compatible tools—all through natural language.

Skip the Cypher Guesswork

You know the frustration: you're deep in a conversation with Claude about your data architecture, then you need to check something in your graph database. Suddenly you're juggling browser tabs, remembering Cypher syntax, and losing your train of thought.

This MCP server suite eliminates that friction entirely. Instead of writing MATCH (p:Product)-[:SOLD_IN]->(o:Order) RETURN p.name, count(o) ORDER BY count(o) DESC LIMIT 10, you just ask: "Show me the top 10 products by order volume."

Four Servers, Four Superpowers

Natural Language to Cypher (mcp-neo4j-cypher)

Your database schema becomes Claude's vocabulary. Ask complex questions about your data and get back proper Cypher queries that actually run. No more "let me check the property names" or "wait, what was the relationship type again?"

Example conversation:

• You: "What are the most connected customers in my network?"
• Claude: Analyzes your schema, generates MATCH (c:Customer) RETURN c.name, size((c)--()) as connections ORDER BY connections DESC LIMIT 5, and shows you the results.

Persistent Knowledge Graph (mcp-neo4j-memory)

Finally, an AI assistant that remembers things across sessions. Store facts, relationships, and insights that persist beyond individual conversations. Your assistant builds up context about your projects, preferences, and work patterns over time.

Perfect for:

• Project knowledge that builds up over weeks
• Personal productivity systems
• Team knowledge sharing
• Research relationship tracking

Aura Cloud Management (mcp-neo4j-cloud-aura-api)

Manage your Neo4j Aura instances without leaving your chat. Scale up for that big data import, spin up a test instance for prototyping, or check your instance health—all from within your existing conversation context.

"Create a new Professional instance called 'ml-experiment' with 8GB RAM and enable Graph Data Science" becomes a single request instead of a five-minute cloud console dance.

Visual Data Modeling (mcp-neo4j-data-modeling)

Design and validate graph data models collaboratively with Claude. Import from Arrows.app, iterate on your schema design, and export back out. Your data modeling sessions become interactive conversations rather than solo drawing sessions.

Real Workflow Impact

Before: Research question → Switch to Neo4j browser → Remember/look up schema → Write Cypher → Debug syntax → Copy results → Switch back to AI tool → Continue conversation

After: "What's the relationship pattern between these entities?" → Immediate answer with visualizations → Continue building on insights

Before: Need test database → Open Aura console → Navigate menus → Fill forms → Wait for provisioning → Copy connection details → Update configs

After: "Spin up a test instance with GDS enabled" → Done in 30 seconds

Getting Started

pip install mcp-neo4j

# Set your Neo4j credentials
export NEO4J_URI=bolt+s://your-instance.databases.neo4j.io:7687
export NEO4J_USERNAME=neo4j
export NEO4J_PASSWORD=your-password

# For Aura management
export NEO4J_CLIENT_ID=your-client-id
export NEO4J_CLIENT_SECRET=your-client-secret

Add to your Claude Desktop config:

{
  "mcpServers": {
    "neo4j-cypher": {
      "command": "python",
      "args": ["-m", "servers.mcp-neo4j-cypher"]
    },
    "neo4j-memory": {
      "command": "python", 
      "args": ["-m", "servers.mcp-neo4j-memory"]
    }
  }
}

Why This Matters

Graph databases excel at relationship analysis, but they've always required specialized knowledge to query effectively. This bridges that gap—you get the analytical power of Neo4j with the conversational interface of modern AI.

Whether you're analyzing social networks, building recommendation engines, or tracking complex business relationships, you can now explore your graph data as naturally as asking a colleague a question.

The 446 GitHub stars and active Neo4j community backing show this isn't just a side project—it's becoming essential infrastructure for teams serious about graph-powered AI workflows.