vector database

A vector database is a specialized data system for storing, indexing, and querying high-dimensional embedding vectors (numerical arrays) so that items with similar meaning or structure lie close together in vector space.

It supports fast nearest-neighbor or similarity search using different techniques, such as graph-based indices, inverted-file clustering, product quantization or hybrid indexes—often in conjunction with metadata filters and keyword + vector hybrid ranking.

Core capabilities of a vector database include insert, update, delete (CRUD/upsert) of vectors and associated metadata, batch ingestion, support for similarity metrics like cosine similarity or inner product, and trade-offs between recall and latency.