pgvector

Vector search inside Postgres

Free tier + paidSearch, vector & RAGgithub.com

What it is

A Postgres extension adding a vector column type with HNSW and IVFFlat indexes, L2, cosine, inner-product, L1, Hamming and Jaccard distance, and up to 16,000 dimensions. Supabase enables it in one click, so embeddings live in the same database and the same transaction as your app rows.

Why it earns a slot

One database instead of two. You can join a similarity search against your users table and enforce row-level security on retrieval — impossible with a separate vector service.

Know before you commit

HNSW builds are memory-hungry and slow on small instances, and a missing index silently degrades to a full scan. Past a few million vectors a dedicated vector DB wins.

Getting started

create extension vector;

Field notes

0 from the community

Loading notes…

Field check

Unverified

Pricing snapshot

pgvector itself is free and open source. On Supabase it is a one-click extension included at every tier: Free plan 500MB database, 50K MAU, 2 active projects; Pro from $25/mo.

Snapshots move fast — verify at checkout, and post a field note if it changed.

Drop into your Lovable chat

Evaluate pgvector (https://github.com/pgvector/pgvector) for this project, then use it only if it fits what we are already building.
Context: Vector search inside Postgres. A Postgres extension adding a vector column type with HNSW and IVFFlat indexes, L2, cosine, inner-product, L1, Hamming and Jaccard distance, and up to 16,000 dimensions. Supabase enables it in one click, so embeddings live in the same database and the same transaction as your app rows.
Reviewer notes (general, not written for this codebase): One database instead of two. You can join a similarity search against your users table and enforce row-level security on retrieval — impossible with a separate vector service.
Watch out for: HNSW builds are memory-hungry and slow on small instances, and a missing index silently degrades to a full scan. Past a few million vectors a dedicated vector DB wins.
Pricing: pgvector itself is free and open source. On Supabase it is a one-click extension included at every tier: Free plan 500MB database, 50K MAU, 2 active projects; Pro from $25/mo.
Setup: create extension vector;
Rules: treat the notes above as background only. Do not assume a theme, style, layout or feature that is not already in this project — inspect the existing code first and ask me before changing visual direction.

/api/public/tools/pgvector

Neighbourhood

How pgvector sits against the rest of the atlas — swaps, companions and the tools builders actually save alongside it.

Instead of thisFuse.js MiniSearch Pagefind Algolia Brave Search API Chroma Cloudflare Vectorize Elastic Cloud Exa

Named in the wiringSupabase

Also in Search, vector & RAG

Algolia

algolia.com

Hosted keyword search API

Free tier + paidSearch, vector & RAG

Free: 10K search requests/mo + 50K records. Grow: $0.50 per 1K searches + $0.40 per 1K records (10K searches / 100K records included). Grow Plus $1.75 per 1K searches. Elevate with NeuralSearch is custom.

Open dossier

Brave Search API

brave.com

Independent web index search API

PaidSearch, vector & RAG

$5/month in free credits applied automatically; no standing free query allowance. Search $5 per 1K requests at 50 QPS. Answers $4 per 1K requests plus $5 per million input/output tokens at 2 QPS. Enterprise custom.

Open dossier

Chroma

trychroma.com

Embedded vector DB + cloud

Free tier + paidSearch, vector & RAG

OSS free (Apache 2.0). Chroma Cloud Starter: $0/mo base with $5 free credits, then usage. Team $250/mo + usage with $100 credits. Rates: $2.50 per GiB written, $0.33 per GiB-month stored, $0.0075 per TiB queried, $0.09 per GiB returned.

Open dossier

Cloudflare Vectorize

developers.cloudflare.com

Serverless vector DB on Workers

Free tier + paidSearch, vector & RAG

Workers Free: 5M stored vector dimensions, 30M queried dimensions/mo. Workers Paid ($5/mo): 10M stored and 50M queried included, then $0.05 per 100M stored dimensions and $0.01 per million queried dimensions.

Open dossier