Vector database
Expanded definition
Vector databases (and vector engines inside Postgres, such as pgvector) index high-dimensional embeddings with approximate nearest-neighbor structures such as HNSW. Operators care about hybrid search, filtering, multi-tenant isolation, re-index cost when the embedding model changes, and recall versus latency. A vector database is storage plus retrieval, not a generator. RAG still needs chunking, reranking, and application-layer permissions.
Related terms
Explore adjacent ideas in the knowledge graph.
Vector database FAQ
What is Vector database?
A vector database stores embedding vectors and runs similarity search, usually with metadata filters and optional hybrid lexical queries.
How is Vector database used in AI systems?
Vector databases (and vector engines inside Postgres, such as pgvector) index high-dimensional embeddings with approximate nearest-neighbor structures such as HNSW. Operators care about hybrid search, filtering, multi-tenant isolation, re-index cost when the embedding model changes, and recall versus latency. A vector database is storage plus retrieval, not a generator.
Related
Comparisons, tools, and models that connect to this idea.