Similar to the estimation used when uploaded a file. Give an estimate to how many tokens are going to be used. Bonus: how many tokens will be left in the current context memory of the selected model.