GraphRAG

GraphRAG models knowledge as an entity and relationship graph (subject-predicate-object triplets).

Two key advantages:

• multi-hop reasoning over relations — user → doctor → hospital → address;
• entity disambiguation — two “Dr. Zhangs” become distinct graph nodes.

Community detection algorithm identifies thematic clusters with summaries.

Limitation: triplet decomposition loses conditional logic and temporal dependencies of natural language.

Recommendation: store full text, keep graph as a specialized index.

Related: RAPTOR, Structured Indexing, User Memory Formats