a way to continue generation when the response is over max tokens.

right now a dev can call complete directly to regenerate. but a function for this with a boolean on the request body for is regeneration or not, would be a nice QOL feature