GraphRAG

About

GraphRAG provides a seamless integration between large language models and a hybrid retrieval system, leveraging the strengths of both graph databases (Neo4j) and vector databases (Qdrant). It enables semantic search through document embeddings, graph-based context expansion following relationships, and hybrid search combining vector similarity with graph relationships, fully integrating with Claude and other LLMs through MCP.

Key Features

• Semantic search using sentence embeddings and Qdrant
• 2 GitHub stars
• Hybrid search combining both approaches
• Full documentation of Neo4j schema and Qdrant collection information
• Graph-based context expansion using Neo4j
• MCP tools and resources for LLM integration

Use Cases

• Hybrid search for information retrieval combining vector similarity and graph relationships
• Integration with LLMs like Claude for enhanced context
• Semantic search through documentation