Vector database
In short: What is Vector database?
A vector database is a data store that holds embeddings and quickly finds the items most similar in meaning to a query.
Explanation
A traditional database searches for exact matches or value ranges. A vector database returns the “nearest neighbours” (the items whose vectors lie closest to the query vector) using specialised indexes that stay fast even across large datasets.
It serves as the standard storage layer for semantic search and RAG applications. Alongside vectors, it stores metadata so results can be filtered by document type, language or access permissions.
Example: every paragraph of a company knowledge base is stored as a vector; when a user asks a question, the database returns the five closest paragraphs to feed into the language model.