All I am going to say is I hope you create a way to ensure that this doesn't just drain all your tokens in one shot, no matter how expensive the model is when I type "Create a prompt to research "topic" It should be more efficient the way that it uses its tokens, please.