🎬 GraphRAG LOCAL: Soluciona PARA SIEMPRE los problemas de tu RAG
Video Details
Channel: Joaquín Ruiz Lite URL: https://www.youtube.com/watch?v=dJ6D_krxpX0 Relevance: ⭐⭐⭐⭐⭐
Summary
Complete tutorial implementing Microsoft’s GraphRAG algorithm locally using Neo4j as the knowledge graph store and Ollama for inference. GraphRAG builds a community graph from documents, enabling multi-hop retrieval that standard vector RAG cannot do. Tutorial covers installation, graph construction, query patterns, and result quality comparison vs standard RAG.
PUMA Relevance
GraphRAG is the Stage 5 extension path for PUMA: instead of retrieving individual similar issues (Stage 4 RAG), GraphRAG can retrieve issue clusters and sprint patterns as graph communities. The Neo4j + Ollama combination is cost-free and reproducible, aligning with PUMA’s Constitution Article 1. Directly applicable to Smart PMO’s sprint planning module.