All of my images are done in SwarmUI with custom prompt commands, wildcards and scripts. Sharing the prompts is pointless so please do not waste your time asking for them. I use only SwarmUI, I loathe ComfyUI, I have extensive (4+ years) experience with Forge/Reforge and SwarmUI as well as python scripting. I can offer some advice for Krea2 though, ignore most "guides", peoples prompts and even AI suggesting the use of tags (realistic, ultra-detailed, HD, 8k, ect) as they only cause Krea2 models to over work and hallucinate trying to do what they do by default. NONE of that is needed, let Krea2 do its job and your images will render MUCH better. If you prefer a more anime or carton style... simply tell the model this is a anime or cartoon image, focus on details, not tags and your image quality will be immensely better. Simple example: it is late at night, the moon is out and shining though an open bedroom window, the room is messy with a bed as the central focus, atop the bed lies a sleeping woman, aged 24 to 40, she is laying on her stomach wearing tight pajamas, the bottom of the pajama set is pulled half down her firm thick ass. the natural moonlight highlights her form.