This is a workflow that I've scrabbled together from other workflows, taking pieces that I like. It's simple enough: It takes the prompt, creates a low resolution image, which is then upscaled by 2, then run through a tagger and optional image recognition to text model to get a more detailed prompt, which is then combined with a controlnet line and depth map to form the positive conditioning for the second model (the photorealistic one), for another pass, then upscaling, and finally detailing.
(Shrugs) That's pretty much it. Nothing special, it's just that someone asked for it on Reddit, and it'd be rude to not be at least a little neighborly.
And, yes, I know it's not that good. It's just my ongoing workflow that I muck about with for personal use. It's not meant to be good :)