I have a stock of 99% IPA already that I use for cleaning the bed of my regular 3D printer, electronics cleaning and for drying things coming out of my ultrasonic cleaner. (I can use salt to drop any water out of IPA, actually.)
With resin 3D printing, my washer uses about a gallon of IPA and it will get super dirty after a while. For that particular case, it's just going to be more efficient to clean the IPA and reuse it until it needs to be distilled. (Wishing the parts in stages will help reduce IPA use, actually. Water washable resin is an option, but I would rather not dump that water down the drain or hassle with hazmat disposal, when applicable.)
Maybe an actual chemist will chime in, but I couldn't find any sources about purifying or recycling IPA, at least none that you could do at home. At a guess, maybe a reverse osmosis system without the finest membrane? Like you said, the water and IPA molecules should be smaller than the dyes and resins.
Maybe also consider washing with the dirty IPA, and just giving a rinse at the end with the clean IPA in order to conserve. Ultimately it sounds like everything I've found is material you're already familiar with, unfortunately.
Thanks for looking around! I already have a good selection of chemistry glass and am no stranger to doing home experiments. (I could probably use a good vacuum filter anyways and maybe I am just searching for an excuse to get one.) Distillation is an option for me, but it'll be last resort. Still, having a few proper distillation pieces would go nice with my collection....
Some filter rigs I have seen are using small RO systems, but that seems like a pain to clean and those people probably need to filter a ton of IPA for it to be cost effective. Regardless, it is absolutely worth exploring more as setting up a filter loop would be awesome until it gets gummed up.
This looked promising as it is dealing with ethanol and plant extracts: https://youtu.be/VjxZVpGv_aM?si=5VFLYQkObCzUawbb .. (This video specifically got me thinking about what could/couldn't be filtered and is the root of this post.)
And absolutely, a multi-stage rinsing system is going to be needed if I scale up, which may be in the near future. Full context: I am studying and planning for a reverse engineering/prototyping business, so there could be a considerable amount of printing in my future and this is particular problem is part of the cost analysis.