I have doubts about how I could solve this use case, my hosting is vercel and the rest api is their native, integrate gpt-4 of openIA with them but without the streaming function or its specific library, just call to the AI with a request to its endpoint, however, the promp is very long and the response takes a long time, maybe 5 min in certain cases, in production vercel obviously gives me a timeout, I saw that if it is via streaming could keep the response open a while. .. however, I don't need to play that response anywhere, I just need it to finish and parse it, to introduce certain data in the database and after the response from the database show a data in the front end, how could I do this?