← All terms
Infrastructure
Vector database
A database optimized for storing embeddings and finding nearest neighbors.
Vector databases index embeddings using approximate nearest-neighbor algorithms (HNSW, IVF) so similarity search stays fast at scale. Popular 2026 choices: pgvector (Postgres extension, default for most), Qdrant, Pinecone, Weaviate, Turbopuffer. Selection criteria: throughput, latency, hybrid-search support, multi-tenancy model, and cost-per-vector at your expected scale.
Related terms
Building with Vector database?
We ship production AI systems built around concepts like this every quarter. Send a brief and get a written proposal in 48 hours.
Send a brief →