Vector search is the process wherein the input query is converted into a vector, represented by numerical embeddings, by the embedded model, and then compared against a vector database to retrieve the most similar results. Vector searches can output text, images, or audio.

Why is vector search important?

Vector search enables more intelligent retrieval and outputs because it understands context, relationships, and semantics. This can provide better recommendation systems. This is unlike traditional keyword searches that rely on exact matches.