pgvector enables storing and querying high-dimensional vectors directly in PostgreSQL. Key features: vector data type (up to 16,000 dimensions), similarity operators (L2 distance, inner product, cosine), and indexing (IVFFlat, HNSW for fast approximate search). Use cases: semantic search (find similar text via embeddings), recommendation systems, image similarity, and RAG applications. Compared to dedicated vector databases: pgvector keeps vectors with your relational data, uses familiar SQL, and avoids managing another system. Indexing is crucial for performance - HNSW provides excellent query speed. pgvector makes PostgreSQL a capable vector database for most AI applications.
🗄️ Database + ⚙️ AI Infrastructure intermediate
pgvector
PostgreSQL extension that adds vector similarity search capabilities for AI and machine learning applications.
</> Related Terms
Vector Database
Specialized database for storing and searching high-dimensional vector embeddings.
Embedding
Numerical vector representation of text, images, or other data for machine learning.
PostgreSQL
Open-source relational database with advanced features like JSONB, full-text search, and extensions.
RAG (Retrieval-Augmented Generation)
AI technique combining vector search with LLMs to provide contextual answers from custom knowledge bases.