Maybe not a feature request: I really tried several and many things to get Bonkers and the Image Generation to not include any text: "do not include any text" or "avoid adding any text", but still minimum 50% of the images created have text :-(