I wonder if they have plans to allow SVG exports in the future. I mean, the file size would probably be ridiculous in a lot of the cases, but for my use case I wouldn't mind it. And sucks about the watermark, maybe they will introduce an option to pay for removing it.
SVG exports would only be meaningful if the model is generating vector images, which are then converted to bitmaps. I highly doubt that's the case, but perhaps someone who has actually looked at the model structure can confirm?
It's just pixels. You can pass them into a tracer
SVG isn't really possible with the model architecture they're using. The diffusion+upscaling step basically outputs 1024x1024 pixels; at no point does the model have a vector representation.
I suppose it's possible that at some point they'll try to make an image -> svg translation model?