Lately the forum’s been flooded with people showing off their images but barely anyone talks about workflows. I feel like that’s putting the cart before the horse. A good image comes from a reproducible pipeline: which base model you pick, how you stack ControlNet, what upscaling chain you use, how you fix faces and hands in post. Just looking at the final output teaches you nothing—change one parameter and everything falls apart.
I’ve maintained a few ComfyUI templates myself: one for e-commerce product shots, one for portrait photography, one for concept scenes. Each template has fixed node connections and samplers, so when requirements change I only tweak a few key nodes. Instead of begging for someone’s prompt for a specific image, we should be breaking down and sharing our workflows. That’s how this community builds real knowledge.
Changing one parameter and the whole thing looks completely different. Just having the prompt is useless—you gotta know how the whole pipeline is set up.
Honestly, asking for prompts is way less useful than breaking down each other’s workflows. That’s how the community actually builds up something real. Anyone else feel the same?