It builds an index suited to high-dimensional similarity search and keeps vectors linked to source content and metadata. At query time it returns candidate records by distance, often with metadata filters and approximate search for speed.
A vector database stores embeddings and retrieves items whose vectors are near a query vector.
It builds an index suited to high-dimensional similarity search and keeps vectors linked to source content and metadata. At query time it returns candidate records by distance, often with metadata filters and approximate search for speed.