HaiAI123

Curated Global AI Tools Directory

Industry update
Flat

Tiger Data open-sources pg_vectorsearch PostgreSQL vector extension

11/19/2026 — 11/19, 13:19·1 sources (all)·1 reports (all)

Key takeaways

  • Tiger Data has released pg_vectorsearch, an Alpha-stage PostgreSQL extension with a PRISM index that uses partitioned compression to cut mem…

Background & analysis

On November 19, 2026, Tiger Data (formerly Timescale, the company behind TimescaleDB) released pg_vectorsearch, an open source PostgreSQL extension for vector similarity search. The project is currently in the Alpha stage. The extension is compatible with pgvector data formats and supports native vector search inside PostgreSQL, so developers can install the extension and create a PRISM index on a vector column without relying on a separate vector database.

The centerpiece of pg_vectorsearch is the PRISM index, which uses a design based on data partitioning and compression rather than the graph-based approach employed by the HNSW index in pgvector. This design reduces memory consumption and shortens index build time. In Tiger Data's own benchmark on 100 million vectors, PRISM built the index on a machine with 128 GB of memory in 1 hour and 40 minutes, while HNSW required 42 hours on a machine with 512 GB of memory.

However, incremental rebalancing has not been fully implemented yet. When the underlying data changes frequently, rebuilding the index may be necessary. Overall, pg_vectorsearch remains at the Alpha stage, with the release and benchmark results reported so far coming from Tiger Data's own testing.

AI-generated from 1 reports · updated 3 days ago

updream
24-hour heatHeat index 25 · peak 33 · 23h ago
24 hours agonow

Reports on this story headlines open the original

Nov 19
  1. Tiger Data has released pg_vectorsearch, an Alpha-stage open-source PostgreSQL extension for vector similarity search. Its PRISM index uses partitioning and compression to cut memory use and build time; in a 100-million-vector test, PRISM built an index in 1 hour 40 minutes on 128GB of RAM, versus 42 hours for HNSW on 512GB. Incremental rebalancing is not fully implemented yet.

    InfoQ 中文AI score 55

Other stories people are talking about

How is heat calculated?About the method

Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.

This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.

Surge
Discussion rising fast
New
First report within 6 hours
Rising
Still gathering discussion

Back to AI industry updates →