Controlling img2img noise amount a possibility?

[mookiexl]

This is an useful feature in the A1111 webUI where it's possible control the amount of noise added during img2img and inpainting. Lowering the amount below the usual 100% results in a simpler image with fewer details, and it can be used for inpainting glitches or cleaning up upscaled images without the AI hallucinating unwanted extra features. I think this would be a good addition to Krita as well.

The usable range in my tests seems to be ~0.8 to 1.0 (but it's situational) Adding more than regular amount is possible but not useful, it just results in ugly artefacts (as shown in the last image). Perhaps it could be presented to the user as a single "simplify" slider (below strength), with zero corresponding to regular 1.0 noise amount and full being either 0, or rescaled to whatever the lowest useful value turns out to be?

I have no idea if this is also an option with ComfyUI or how difficult it would be to implement, sorry if I suggested something impossible.

EDIT: it also seems to work the best with Euler A, where it effectively blurs out anything not described in the prompt but leaves the subject mostly intact.