Meet LEANN: An Open-Source Vector Index that Shrinks a 60M-Chunk RAG Index from 201 GB to 6 GB

UC Berkeley's Sky Computing Lab developed LEANN, an open-source vector index that reduces storage for a 60M-chunk Wikipedia index from 201 GB to 6 GB by recomputing embeddings at query time instead of storing them. LEANN matches traditional HNSW accuracy while drastically cutting storage needs and runs efficiently on laptops and workstations.