In the chat UI, gpt can remember / recall previous messages / prompts / responses. Is there a way to do this with with the API as well? I mean, get it to look through the entire conversation before answering a prompt? i can do this programatically (like load the previous messages in the prompt) but eventually they get too much - too many tokens…
This is the only way to do it at this time. One thing you can try is to summarize super-old messages in the chain, but you begin to lose coherence a bit after a while. Hope this helps.