Short Definition
A vector database stores and searches embeddings using similarity rather than exact keyword matching.
Intuition
It helps find items that mean similar things, even when they do not share the same words.
Technical Definition
A vector database indexes high-dimensional vectors and supports nearest-neighbor search using metrics such as cosine similarity, dot product, or Euclidean distance.
Example
A knowledge-base search system stores document embeddings in a vector database for semantic retrieval.
Common Misunderstandings
A vector database does not replace all traditional databases.
Vector search still needs chunking, metadata filters, and evaluation.