So if you were doing semantic search for eg, you would have to first take the query that the user typed in, send it to OpenAI, get the embeddings from their API, then send that to Typesense using the

vector_query

parameter