CargoCub's Art Gallery

If I'm not interested in seeing what you request, I'll let you know but I won't make it.

So I usually start with a vague prompt with a low step count but a larger number of images produced. Something like "A shirtless dadbod wearing a speedo in public." I throw in some negative tags like "abs, thin, twink." From there I select and refine an image. Sometimes I do an image-to-image and others it's inpainting to get something right. Sometimes I'll inpaint to get a particular detail right and then img-to-img to make it all cohesive. I almost always add body hair last as it can mess up the chest and belly and look off.

I can try, that slick look is hard for the ai to make look real.