mcp-knowledge-graph by moche-ai - MCP Server

MCP Knowledge Graph

A Model Context Protocol (MCP) server that provides verified knowledge with trust scores, reasoning capabilities, and fact-checking.

🌐 Live Server: https://mcp.moche.ai

Why Use This?

Unlike simple document search (like Context7), this MCP server provides:

Feature	Simple Search	MCP Knowledge Graph
Document lookup	✅	✅
Trust scores	❌	✅ Verified information
Relationships	❌	✅ Dependencies, alternatives
Reasoning	❌	✅ "Why" and "How" answers
Best practices	❌	✅ Pitfalls and recommendations

Quick Start

Use the Public Server

Add to your MCP client configuration:

{
  "mcpServers": {
    "knowledge-graph": {
      "transport": {
        "type": "sse",
        "url": "https://mcp.moche.ai/mcp/sse"
      }
    }
  }
}

Self-Host with Docker

# Clone the repository
git clone https://github.com/moche-ai/mcp-knowledge-graph.git
cd mcp-knowledge-graph

# Set environment variables
export NEO4J_PASSWORD=your_password

# Run with Docker Compose
docker compose up -d

Install from Source

pip install mcp-knowledge-graph

# Or install from source
pip install -e .

# Run the server
mcp-kg-server

Available Tools

Tool	Description
`search_knowledge`	Search for verified information with trust scores
`get_context`	Get complete context: dependencies, alternatives, integrations
`get_dependencies`	Get dependency chain for installation order
`get_alternatives`	Find alternative technologies/tools
`get_best_practices`	Get recommendations and pitfall warnings
`get_stats`	Get knowledge graph statistics
`infer_relation`	Infer relationships between two technologies
`find_path`	Find connection paths between concepts
`recommend`	Get technology recommendations (similar, alternative, complement)
`find_similar`	Find similar technologies by category/tags

Usage Examples

Search Knowledge

curl -X POST https://mcp.moche.ai/mcp/tools/call \
  -H "Content-Type: application/json" \
  -d '{
    "name": "search_knowledge",
    "arguments": {"query": "vector database", "min_trust": 0.8}
  }'

Get Context

curl https://mcp.moche.ai/knowledge/context/langchain

Get Dependencies

curl -X POST https://mcp.moche.ai/mcp/tools/call \
  -H "Content-Type: application/json" \
  -d '{
    "name": "get_dependencies",
    "arguments": {"name": "langgraph", "max_depth": 3}
  }'

API Endpoints

Endpoint	Method	Description
`/health`	GET	Health check
`/mcp/info`	GET	MCP server info
`/mcp/tools/list`	GET	List available tools
`/mcp/tools/call`	POST	Call a tool
`/mcp/resources/list`	GET	List resources
`/mcp/resources/read`	POST	Read a resource
`/mcp/sse`	GET	SSE streaming endpoint
`/knowledge/search`	GET	REST search endpoint
`/knowledge/context/{name}`	GET	REST context endpoint
`/knowledge/stats`	GET	Graph statistics
`/knowledge/infer/relation`	GET	Infer relations between technologies
`/knowledge/infer/path`	GET	Find paths between concepts
`/knowledge/recommend/{name}`	GET	Get recommendations
`/knowledge/similar/{name}`	GET	Find similar technologies
`/knowledge/viewer`	GET	3D Knowledge Graph Viewer
`/docs`	GET	OpenAPI documentation

Configuration

Environment Variable	Default	Description
`MCP_PORT`	8780	Server port
`NEO4J_URI`	bolt://localhost:7687	Neo4j connection URI
`NEO4J_USER`	neo4j	Neo4j username
`NEO4J_PASSWORD`	-	Neo4j password (required)
`MIN_TRUST_SCORE`	0.7	Default minimum trust score

Trust Scores

All information has a trust score (0-1):

Level	Score	Meaning
Verified	0.9-1.0	From official sources (GitHub, official docs)
High	0.7-0.9	From trusted sources (ArXiv, reputable blogs)
Medium	0.5-0.7	Community-verified
Low	0.3-0.5	Single source, unverified
Unverified	0.0-0.3	Needs verification

Contributing

Contributions welcome! Please:

Fork the repository
Create a feature branch
Submit a pull request

License

MIT License - see for details.

Support

If you find this project helpful, consider supporting its development:

⭐ Star this repository
🐛 Report issues or suggest features
💬 Share with others

ETH: 0x0000C4CE7f73886Cce8C43d4b481f0cBEbAC0000