Vector store

A vector store, or vector database, holds embeddings, the numeric fingerprints of meaning, and is built for one job: given a new piece of text, find the stored items whose meaning is closest. That similarity search is what a RAG system runs every time it needs to pull relevant passages.

It is the right home for meaning-based lookup over writing. It does not replace a normal database, because exact filters, totals, and calculations over structured records still belong in regular tables queried directly.