Hi @Jason Bosco I am also doing a hybrid search using my own embeddings and am hoping I can get your input on an error. I'm using all-mpnet-base-v2 to create a 768 dimensional vector which is stored in a variable _*my_vec.*_ My query is as follows:

search_parameters = {

'q' : 'harry potter',

'query_by': 'title,text_embedding',

'vector_query: f'text_embedding:({my_vec}, k:100)',

}

response = client.collections[typesense_collection].documents.search(search_parameters)

I'm encountering this error.

[Error 400] Query string exceeds max allowed length of 4000. Use the multi_search end-point for larger payloads.

When I attempt to use the multi_search it returns two separate results as expected but they are not combined like the normal search above. Do I need to move to auto-embeddings or do I need to decrease the vector to something smaller such as all-MiniLM-L6-v2?