A neurobiologically inspired long-term memory framework that builds a knowledge graph over documents and retrieves with Personalized PageRank, enabling continual integration of knowledge. HippoRAG 2 improves multi-hop associativity and sense-making.
- Storage
- Open knowledge graph built from extracted entities/relations over a document corpus, plus passage embeddings.
- Retrieval
- Personalized PageRank over the knowledge graph for single-step multi-hop retrieval; HippoRAG 2 adds deeper passage integration and online LLM use.
- Self-host
- Self-host: moderate
- License
- MIT
- Pricing
- Open source (MIT), free to self-host · Free / OSS
- GitHub stars
- 3,797
- Last release
- 2025-02-27
- Last commit
- 2025-09-04
- First catalogued
- 2026-06-28
Strengths
- Peer-reviewed (NeurIPS'24)
- Single-step multi-hop retrieval via Personalized PageRank
- More efficient offline indexing than GraphRAG/RAPTOR/LightRAG per the authors
Watch out
- Primarily a research artifact; last repo push ~Sept 2025, so maintenance cadence is uncertain
- Targets corpus QA / continual knowledge integration more than conversational user memory
Best for
- Multi-hop retrieval and knowledge integration over large document corpora
Benchmark results
No sourced results yet.
Sources
- https://github.com/OSU-NLP-Group/HippoRAG (vendor)
- HippoRAG: Neurobiologically Inspired Long-Term Memory for LLMs (paper)
- From RAG to Memory: Non-Parametric Continual Learning for LLMs (HippoRAG 2) (paper)
Last verified 2026-06-28 · updated by refresh-framework-cards