← all repositories

timescale/pgvectorscale

A PostgreSQL extension providing high-performance vector search using the DiskANN algorithm for AI embedding storage and retrieval.

pgvectorscale
Velocity · 7d
+2.8
★ / day
Trend
steady
star history

pgvectorscale extends PostgreSQL with a new StreamingDiskANN index type and Statistical Binary Quantization for efficient embedding storage and approximate nearest neighbor search. It complements pgvector by providing higher throughput and lower latency for AI workloads, supporting label-based filtered vector search for precise retrieval. Built in Rust using the PGRX framework, it targets large-scale embedding storage in AI applications.

heatdrop uses Google Analytics to see which pages get read — nothing else. Your call. How we handle data.