Hi HN! Curious if there are any platforms that allow users to upload their Stable Diffusion workflows so that others can easily generate images from them while on the platform's hardware?<p>There are some helpful communities for sharing ComfyUI workflow JSONs for example, but it’d be nice to be able to pay a small fee to run it hosted elsewhere with one click (plus whatever other inputs), instead of copying it into RunDiffusion, RunPod, or my local ComfyUI and dealing with the slowness and missing custom nodes/models. I know it comes with the cost of customization, but think it could also help bridge the gap for folks less deep in the weeds (e.g. many midjourney users) to try out custom SD workflows. Even though MJ is great, the total sum of community-driven SD custom checkpoints, LoRAs, VAEs, etc is surely more powerful than MJ alone!<p>Also curious about this from the creator side since it’d be sweet to monetize workflows if there's enough consumer demand once the compute is paid for
I don't think so, you'll have to build it yourself.<p>In the meantime, for dealing with missing custom nodes/models you can use ComfyUI Manager <a href="https://github.com/ltdrdata/ComfyUI-Manager">https://github.com/ltdrdata/ComfyUI-Manager</a> which doesn't fully automate this away, but does make it easy to solve with just a few clicks.
i think magespace maybe, i tried prerendering and selling as nfts which was a total waste of time.<p>I'm running <a href="https://ebank.nz" rel="nofollow">https://ebank.nz</a> an AI Art Generator thats also open sourced.<p>even myself i'm thinking of moving to opendalle, can't beat open source...<p>would like to work with you on that if your interested its open sourced at Netwrck/stable-diffusion-server which would serve as a good base for a project like that, Needs to do model swapping caching for this kind of use case perhaps.
Or a comfyUI approach is good too.