Vector database
Also: vector store
A store built to find the vectors nearest a query vector quickly, across millions of items, usually alongside metadata to filter on.
What people get wrong
Below roughly a million vectors, a Postgres extension or even a numpy array in memory is fast enough. A dedicated database is an operations decision, not a correctness one.
A definition is not an understanding. This is taught properly in Storing and searching vectors, part of Vector databases.
Learn this properlyTerms from the same topic