Report card

Do AI agents pick pgvector?

How often agents choose pgvector when a developer needs vector database — measured across Claude, GPT, Gemini.

pgvector

#2 of 5 vector db
13%Pick Rate
95% CI 9%18%

When developers ask an AI agent for vector db, pgvector is picked 13% of the time — ranking #2 of 5 measured across Claude, GPT, Gemini.

Picked 13% Named, not picked 24% Never surfaced 63%

Beaten by Pinecone (30%).

Awareness

0 docs in FineWeb (10BT sample) · 1 URLs in Common Crawl

In the raw web crawl, but thin in the filtered corpus models actually train on — an awareness gap at the source.

AdoptionDid the pick convert to installs? Public Adoption = npm downloads of your packages. The proof half of the funnel.

1.5M installs / month ▲ 11% MoM

Growing despite a low Pick Rate — winning the agent pick is upside, not a dependency.

Get the full report

The per-model and per-surface breakdown for pgvector — where it wins, where it loses, and to whom — plus an alert when the Pick Rate moves.

About pgvector

A PostgreSQL extension that adds vector storage and similarity search to your database.

pgvector lets you store vector embeddings directly in PostgreSQL and query them using distance metrics like L2, cosine, and inner product. It supports approximate nearest-neighbor indexes (HNSW and IVFFlat) for fast similarity search at scale. It's aimed at developers building search, recommendation, or AI-backed features who want to keep vector data alongside relational data in Postgres.

Where to start

Install the pgvector extension in your PostgreSQL database, then use the pgvector-node package with your existing Node.js database library (node-postgres, Prisma, Drizzle, Sequelize, and others are supported) to store and query vectors.

Install

Links and summary verified from public sources.

The rest of the Vector Database ranking

← Full Vector Database leaderboard · How this is measured