It seems in latest (or one of the latest updates) the little button that allowed you to switch between text and image generation has disappeared. I am trying to do a roleplay, which is hindered by the AI interpreting many of the prompts as images, rather than me describing the setup of a scene.