All Tools
H
RAGFreeOpen Source
HIPPORAG
Neurobiologically inspired long-term memory for large language models
MIT
ABOUT
Traditional RAG systems struggle with multi-hop reasoning and continual knowledge integration across large document collections. HippoRAG solves this by modeling memory as a dynamic knowledge graph with personalized PageRank, allowing LLMs to perform complex sense-making and retrieve facts across disconnected documents.
INSTALL
conda create -n hipporag python=3.10
conda activate hipporag
pip install hipporag
INTEGRATION GUIDE
1. Multi-hop question answering over large, heterogeneous document corpora
2. Continual knowledge integration into a live, evolving memory graph for long-term learning
3. Complex narrative sense-making and factual memory retrieval across disconnected sources
4. Local and private vLLM deployment for sensitive enterprise knowledge bases
5. Dynamic knowledge base construction and maintenance from streaming documents
TAGS
ragknowledge-graphmulti-hop-qallm-memoryneuripspage-rankcontinual-learningvllmpython