pgvector
Postgres extension that stores vectors and runs exact or approximate similarity search inside the database.
How the 8.4 was reached
- Quality of output8/10
23,194 GitHub stars, active with recent pushes and 17 open issues; supports exact and approximate search with several distance metrics. github.com
- Ease of use8/10
Available through Docker, Homebrew, APT, Yum and more, and preinstalled with Postgres.app; README docs cover install. github.com
- Pricing value10/10
Completely free open-source software under the PostgreSQL licence, with no software cost. costbench.com
- API and integration quality8/10
Works with any language that has a Postgres client and supports several vector types; documented in the README. github.com
- Solves the problem it claims to solve8/10
Vector search inside Postgres with ACID, JOINs and point-in-time recovery, backed by strong GitHub adoption. github.com
8.4/10 is the mean of the 5 criteria that apply. Scored on the five-criteria rubric v1, 29 September 2026. How scoring works
Pricing
- PostgreSQL licence
- Supports Postgres 13+
Key Features
- Exact and approximate nearest neighbor search
- L2, inner product, cosine, L1, Hamming and Jaccard distances
- Single-precision, half-precision, binary and sparse vectors
- ACID compliance, JOINs and point-in-time recovery
- Install via Docker, Homebrew, APT, Yum, PGXN and more
Pros & Cons
Pros
- Vectors live beside the rest of your data in Postgres
- No software cost
- Works with any language that has a Postgres client
- Active project with 23,194 stars and 17 open issues
Cons
- Docs are the README, with no separate reference site
- Requires running and tuning Postgres yourself, or on a host that supports it
- Dedicated vector databases may offer more purpose-built features
pgvector is a sensible pick for teams already on Postgres who want vector search without adding another database. Those needing a managed, vector-only service should compare Pinecone or Qdrant.
Try pgvector →Added to scored.tools on
Competitors to pgvector
Other tools in the database category worth comparing.
Pinecone
7.4/10Vector database for building knowledgeable AI applications with fast similarity search
Qdrant
8.2/10High-performance open-source vector database written in Rust for AI applications.
Chroma
7.6/10Open-source vector database and search infrastructure for AI applications
Weaviate
6.6/10Open-source vector database designed for AI applications and semantic search
More Articles Featuring pgvector
Best Open Source Vector Databases for Production RAG 2026
Honest comparison of the top open source vector databases for production RAG in 2026, ranked by performance, scalability, and operational maturity.
roundupBest Open Source RAG and Vector Search Tools 2026
An honest ranking of the open source RAG frameworks and vector databases actually worth running in production in 2026, based on ingest speed, recall, and operational cost.
roundupBest AI Database Platforms 2026: Vector & Hybrid Search Ranked
We compared the leading AI databases on ingestion speed, hybrid search quality, scaling, and price. Here's how Pinecone, Weaviate, Qdrant, Chroma, Milvus, pgvector, LanceDB, and Turbopuffer actually compare.
roundupBest Vector Databases 2026: Top 8 Tools for AI Applications
Complete comparison of the top vector databases for AI applications, ranked by performance, features, and real-world usability.