Hey this looks kinda familiar 
I know thereās already been a lot of awesome feedback here and suggestions, but I did want to speak up as someone who loves the Pamās -> Scales match up as a sequencing solution.
I didnāt like Marbles (too variable) or Ansible (not variable enough). With Pam you have a variety of shapes you can attenuate, phase differently, manipulate with CV, and feed into scales, which you can āplayā by selecting the notes you want sequenced as the song progresses.
(On top of that Scales can be a vanilla step sequencer if you want that sometimes.)
Assuming youāre not self-patching Pamās for the variability, you can sequence two voices with three channels out of Pamās, leaving 5 for other modulation duties.
Anyway! Part of the fun of modular is finding the combos that work for you. Just wanted to stick up for this approach some 