Just describe a scene, no matter how crazy it may be. MarKI then generates the right image for you within a few seconds. MarKI now uses the "Stable Diffusion" model, which performs its calculations on 128 super-fast A100 graphics cards in a few seconds.
