Pinecone releases open-source pg_pinecone extension for Postgres
Apache 2.0-licensed pg_pinecone lets teams choose a local pgvector, native or Pinecone Serverless search engine for existing pgvector columns while keeping the same SQL queries.
AT A GLANCE
Key points
- One pgvector column can use pgvector, native or remote search engines.
- Native search runs in Postgres without account or network access.
- The extension supports BM25, vector and hybrid search.
What is changing?
Pinecone released pg_pinecone on October 27, an open-source extension that brings vector search into PostgreSQL. According to the official GitHub repository, it can index existing pgvector columns without changing them and offers three engines per table: pgvector, the default native engine and a remote Pinecone Serverless engine.
The native engine runs search inside Postgres and can apply filters during retrieval without requiring an account or network access. The remote engine moves search to a Pinecone Serverless index while rows remain in Postgres. Engine switches are reversible while preserving the same query surface.
The extension also supports BM25 full-text, vector and hybrid search through a Pinecone index kept in sync with a Postgres table. The quickstart requires PostgreSQL 17 or 18, server superuser access and pgvector, and the project is released under Apache 2.0.
Explore the original sources
Source date: 27 October 2026.
Published by AI Radar: 27 Oct 2026, 17:15 (Istanbul, UTC+3).
FOLLOW THE TOPIC
Topics and tags
Share your discovery.
Send this story to a friend.
Reader comments
What do you think about this development? Share your experience or question. Comments are published after editorial approval.
Your name/nickname and comment become public if approved. Read: Privacy notice.
Loading comments…