As per my comment on Youtube: Of course the whole VST hassle normally comes with the territory but that is actually my main point: Must one really add a completely new kitchen wing to the house to get that one pefect expresso machine whistling? :-)
A semi-standalone situation as you (seem to) suggest could however be just the right ticket for me. I can then keep the rest of my workflow in the good old "hardware only" realm and just let the laptop do all the number crunching on one instance of Kaivo, which seems to be renowned for being processor-hungry anyway. So if I could run it as a single semi-standalone polyphonic poly aftertouch sensitive synth with a reasonably low latency I would already be more then happy since my preffered working method is based on recording real time performances anyway.
One extra remark: An alternative to full poly aftertouch processing would be to at least have Kaivo available in a multitmbral mode, since somebody has already cracked the poly atertouch to multichannel /-timbral nut for me, assuming that would not defeat the whole K.I.S.S. principle again.
Anyway: If you could point out a sure fire method to achieve such a solution I would appreciate it very much.
A setup with a single instance in a minimal DAW seems like a good solution for you. DAWs do not take too much CPU, generally. They add complexity but you can set this up once and forget about it.
Reaper is a good option and inexpensive.
A newcomer on the scene is called Gig Performer. They are trying to support just the use you envision. I have no experience with it myself yet.
To get a multitimbral use I would make two instances of Kaivo on different tracks and use each on a different MIDI channel.
Sorry to be stubborn but that is actually exactly what I am trying to avoid, now is it? Are there any experiences with hosts, for instance, Savihost?