PostgreSQL 17 with pgvector on Ubuntu 24.04 LTS

Overview

This appliance pairs the PostgreSQL relational database with the pgvector extension, turning a proven SQL engine into a first class vector store. Alongside ordinary tables you can store high dimensional embeddings in a dedicated vector column, build approximate nearest neighbour indexes, and run similarity queries that rank rows by distance to a query vector. Because it is just PostgreSQL, you keep transactions, joins, SQL and the entire driver and tooling ecosystem while adding semantic capabilities.

It suits retrieval augmented generation over your own documents, semantic and hybrid search, recommendation and deduplication, and any application that needs to combine relational data with similarity over embeddings, all on CPU with no GPU required.

Common uses

• Retrieval augmented generation and semantic search over your own documents
• Storing and querying embeddings next to relational data with plain SQL
• Recommendation, deduplication and similarity ranking on CPU