All Tools
P
Vector DBFreeOpen Source
PGVECTO.RS
Vector search extension for PostgreSQL
Apache-2.0
ABOUT
Teams using PostgreSQL for their application database need to add vector similarity search for RAG and embedding-based retrieval without running a separate vector database service — but existing PostgreSQL vector extensions either lack ANN indexing (pgvector) or are written in C with limited filter support. pgvecto.rs is a Rust-based PostgreSQL extension that implements high-performance HNSW and IVF-PQ indexes with full SQL integration, pre-filtering on scalar columns, disk-based index storage that doesn't require indexes to fit in RAM, and distance functions (cosine, Euclidean, dot product, Hamming) — enabling hybrid structured + vector queries within PostgreSQL without additional infrastructure.
INTEGRATION GUIDE
1. Add vector similarity search to existing PostgreSQL databases for RAG retrieval, enabling semantic search over document embeddings with pre-filtering on metadata fields like date, author, and category
2. Build hybrid search applications that combine full-text PostgreSQL queries with vector ANN search, ranking results by combined BM25 and semantic similarity scores in a single SQL query
3. Deploy production vector search with disk-based HNSW indexes that scale to billions of vectors without requiring all indexes to fit in RAM, reducing infrastructure costs for large-scale embedding retrieval
TAGS
postgresqlvector-searchrustannextension