NetworkX MCP Server

The first NetworkX integration for Model Context Protocol - Bringing graph analysis directly into your AI conversations.

🚀 What is this?

NetworkX MCP Server enables Large Language Models (like Claude) to perform graph analysis operations directly within conversations. No more context switching between tools - analyze networks, find communities, calculate centrality, and visualize graphs all through natural language.

🎯 Key Features

• 13 Essential Graph Operations: From basic graph creation to advanced algorithms like PageRank and community detection
• Visualization: Generate graph visualizations on-demand with multiple layout options
• Import/Export: Load graphs from CSV, export to JSON
• Zero Setup: Works immediately with Claude Desktop or any MCP-compatible client
• First of Its Kind: The first NetworkX server in the MCP ecosystem

🌟 Why NetworkX MCP Server?

• Natural Language Graph Analysis: Describe what you want to analyze in plain English
• No Database Required: Unlike graph database integrations, this works with in-memory graphs
• Instant Insights: Get centrality metrics, find communities, and discover patterns immediately
• Visual Understanding: See your graphs, don't just analyze them
• Enterprise Ready: Production-grade security, monitoring, and scale (Enterprise Edition)

📊 Editions

Community Edition (Free)

• 13 Graph Operations: Complete NetworkX functionality
• Visualization: PNG output with multiple layouts
• Import/Export: CSV and JSON support
• Zero Setup: Works with Claude Desktop immediately

pip install networkx-mcp-server

🏢 Enterprise Edition

• Everything in Community Edition +
• 🔐 Enterprise Security: OAuth 2.1, API keys, RBAC
• ⚡ Rate Limiting: Per-user and per-operation quotas
• 📊 Monitoring: Prometheus metrics, audit logging
• 🛡️ Input Validation: Comprehensive security validation
• 📈 Resource Control: Memory and execution limits
• 🚀 Production Ready: Health checks, Docker support

pip install networkx-mcp-server[enterprise]

📖 Enterprise Guide | Demo | Security

📊 Available Operations

Core Operations

• create_graph - Create directed or undirected graphs
• add_nodes - Add nodes to your graph
• add_edges - Connect nodes with edges
• get_info - Get basic graph statistics
• shortest_path - Find optimal paths between nodes

Analysis Operations

• degree_centrality - Find the most connected nodes
• betweenness_centrality - Identify bridges and key connectors
• pagerank - Google's PageRank algorithm for node importance
• connected_components - Find isolated subgraphs
• community_detection - Discover natural groupings

Visualization & I/O

• visualize_graph - Create PNG visualizations with multiple layouts
• import_csv - Load graphs from edge lists
• export_json - Export graphs in standard formats

🚦 Quick Start

Community Edition

# Install community edition
pip install networkx-mcp-server

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "networkx": {
      "command": "networkx-mcp",
      "args": []
    }
  }
}

Enterprise Edition

# Install enterprise edition with security features
pip install networkx-mcp-server[enterprise]

# Set up authentication (generate secure key)
export NETWORKX_MCP_SECURITY_API_KEYS="$(python -c 'import secrets; print(secrets.token_urlsafe(32))')"

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "networkx-enterprise": {
      "command": "networkx-mcp-enterprise",
      "args": [],
      "env": {
        "NETWORKX_MCP_SECURITY_API_KEYS": "your-secure-api-key-here",
        "NETWORKX_MCP_RATE_LIMIT_ENABLED": "true",
        "NETWORKX_MCP_MONITORING_METRICS_ENABLED": "true"
      }
    }
  }
}

📖 Complete Enterprise Setup Guide

Basic Usage Example

Human: Create a social network graph and analyze it

Claude: I'll help you create and analyze a social network graph.

[Creates graph "social_network"]
[Adds nodes: Alice, Bob, Charlie, David, Eve]
[Adds edges: Alice-Bob, Alice-Charlie, Bob-Charlie, Bob-David, Charlie-Eve]

Now let me analyze this network:

[Calculates degree centrality]
Most connected: Bob and Charlie (degree centrality: 0.75)

[Calculates betweenness centrality]
Key bridges: Charlie (betweenness: 0.58) - connects different parts of the network

[Detects communities]
Found 2 communities: [Alice, Bob, Charlie] and [David, Eve]

[Visualizes the graph]
Here's what your social network looks like: [displays graph image]

🎓 Example Use Cases

1. Social Network Analysis

• Identify influencers and key connectors
• Find communities and cliques
• Analyze information flow patterns

2. Transportation Planning

• Find shortest routes between locations
• Identify critical intersections
• Analyze network resilience

3. Knowledge Graphs

• Map concept relationships
• Find learning paths
• Identify prerequisite chains

See the demos/ folder for complete examples.

📈 Performance

• Memory: ~70MB (including Python, NetworkX, and visualization)
• Graph Size: Tested up to 10,000 nodes
• Operations: Most complete in milliseconds
• Visualization: 1-2 seconds for complex graphs

🛠️ Development

Running from Source

# Clone the repository
git clone https://github.com/Bright-L01/networkx-mcp-server
cd networkx-mcp-server

# Install dependencies
pip install -e .

# Run the server
python -m networkx_mcp.server_minimal

Running Tests

pytest tests/working/

📚 Documentation

• API Reference - Detailed operation descriptions
• Examples - Real-world use cases
• Contributing - How to contribute

🤝 Contributing

We welcome contributions! This is the first NetworkX MCP server, and there's lots of room for improvement:

• Add more graph algorithms
• Improve visualization options
• Add graph file format support
• Optimize performance
• Write more examples

📄 License

MIT License - See LICENSE for details.

🙏 Acknowledgments

• NetworkX - The amazing graph library that powers this server
• Anthropic - For creating the Model Context Protocol
• The MCP community - For inspiration and examples

---

Built with ❤️ for the AI and Graph Analysis communities