With vector search, every item in the dataset is technically a "match" to some level of closeness. The

k

parameter determines the top number of items to consider for the result set (regardless of distance). And the

distance_threshold

sets a defined value for how close you want the fetched results to be. You can use both in combination