How about creating a "sub-version" which is just a single, morphable drum voice with no sequencer. Great for simplified use in (ableton) drum racks, (bitwig) drum machines, (logic pro) drum machine designer, (maschine) groups, etc. Simplified routing, better for assigning tweakable parameters per voice, lower CPU cost.

You can do most of what I want this for now of course just by using an entire instance of microtonic for each drum-rack channel/voice, but it seems cpu-wasteful. Unless it actually isn't... Do muted or never-triggered microtonic drum voices consume any CPU dsp overhead? If there's actually no or very low computational cost in unused channels then I guess this would be unnecessary.

Sorry for the absurdly late reply, this one somehow got buried back in January.

Unused voices do not take any CPU, so the overhead in that kind of setup is mainly per plug-in instance. For that reason, Microtonic Multi is usually the better fit if you want one channel per pad/voice with separate routing and processing.

For an example of that approach in Ableton, see this thread:
https://soniccharge.com/forum/topic/71-microtonic-drum-rack-for-ableton-live

This has been mused about before of course: https://soniccharge.com/forum/topic/2790-minimicrotonic-a-single-drumpatch-device

I extensively use AbletonDrumRacks and it would be so convenient and neat to just drag single voice instances of microtonic into it.