Comment by diziet

I would also love to see a token budget use for the chats -- to know when the model is about to run out of context. It's crazy this is not there.