SAG
SAG is a SQL-driven Retrieval-Augmented Generation (RAG) engine designed to automatically build knowledge graphs during the query process rather than pre-building them. It transforms raw text into semantic atomic events and extracts multi-dimensional natural language vectors for each event. At query time, SAG dynamically constructs relationship networks by combining precise SQL retrieval, vector semantic search, and PageRank algorithms in a three-stage process of recall, expansion, and reranking. This event-centric approach eliminates the need to maintain large static graphs, significantly reducing maintenance costs while improving retrieval accuracy and scalability compared to traditional RAG and static GraphRAG systems. Every result provides complete traceability to its source and connection pathways. The system supports custom entity types to fit various business scenarios. SAG is built on Python, FastAPI, and Next.js, making it suitable for developers seeking an easy-to-degenerate local or enterprise RAG