Wow. Yeah, that would be amazing. vid2vid has a non-commercial license so it is probably not something Team Vuo could add. But if anyone else would be interested in coding it, you can embed Python in C/C++.
Thanks @Kewl, it's really the inside Vuo stuff that I'm thinking about, how one would capture a 3d scene to wrap a cube without weird distortions and misalignments.
Would you build a kind of 6 camera setup? and how would that stitch together? What if the room wasn't perfectly cubed? Could we maybe stitch two 180degree fisheye cameras together and somehow use that to build the texture map?