# NeuroSpace MCP Server

An MCP (Model Context Protocol) server that exposes a Neo4j knowledge graph to AI assistants like Claude.

## Overview

This server provides tools for AI models to query and explore a Neo4j-based knowledge graph using the standardized MCP protocol.

## Tools

| Tool | Description |
|------|-------------|
| `search_entities(name)` | Fuzzy case-insensitive search for nodes by name |
| `get_neighbors(entity_name)` | Returns all connected nodes + relationship types |
| `run_cypher_readonly(query)` | Executes arbitrary read-only Cypher queries |

## Setup

```bash
python -m venv venv
.\venv\Scripts\Activate.ps1  # Windows
pip install -r requirements.txt
```

## Environment Variables

| Variable | Default | Description |
|----------|---------|-------------|
| `NEO4J_URI` | `bolt://localhost:7687` | Neo4j Bolt connection URI |
| `NEO4J_USER` | `neo4j` | Neo4j username |
| `NEO4J_PASSWORD` | `password` | Neo4j password |

## Testing with MCP Inspector

```bash
mcp dev server.py
```

This opens a browser-based Inspector UI where you can:
- See all 3 tools auto-discovered from the `@mcp.tool()` decorators
- Call `search_entities` with a name to fuzzy search
- Call `get_neighbors` to explore graph connections
- Test `run_cypher_readonly` with safe MATCH queries

## Claude Desktop Integration

Copy `claude_desktop_config.example.json` to your Claude Desktop config directory and update the Neo4j credentials.