Vector Database.
A database optimized for storing and searching embedding vectors, enabling fast semantic similarity search across millions of text chunks or documents.
A vector database stores high-dimensional numerical vectors (embeddings) and provides efficient similarity search — given a query vector, find the most similar vectors in the database. This is the core infrastructure component of RAG systems, powering semantic search over private document collections, knowledge bases, and long-term memory for AI agents.
Popular vector databases include Pinecone, Weaviate, Qdrant, and Milvus (purpose-built), as well as pgvector (a PostgreSQL extension) and Chroma (lightweight and developer-friendly). They use approximate nearest neighbor (ANN) algorithms like HNSW and IVF to perform billion-scale searches in milliseconds.
The key metric for vector databases is recall at K — what fraction of the true top-K nearest neighbors are returned by the approximate search. Most production systems tune for a balance of recall and latency. As models improve and RAG architectures mature, vector databases have become core infrastructure for enterprise AI applications.