Hi Neo4j community!
I’m already leveraging Neo4j as my main database and looking to maximize its capabilities for Retrieval-Augmented Generation (GraphRAG) with LLMs. What are the different patterns, architectures, or workflows available to build or convert a solution to “GraphRAG” with Neo4j as the core knowledge source? @alexander.erdl
If you have neo4j as a knowledge source already you could use fulltext or vector retrieval on the nodes or hybrid retrieval, or agentic workflows. You could start here: User Guide: RAG — neo4j-graphrag-python documentation