Im in the process of doing something similar : http://community.axoloti.com/t/axocube-project
but Im not using the PI (even the PI2 I have) for patching, I see very little advantage to that, over a (even cheap) laptop, it won't be a nice experience. Instead I consider the PI as part of a 'performance setup' (at least initially) gluing things together that the Axoloti cannot currently do e.g. control multiple axolotis via usb hubs, other devices.
the low level USB protocol of axoloti could be used to still get the 'live' elements of the axoloti UI, Im actually reconsidering this approach at the moment with the Push 2 e.g. rather than driving the Push2 directly from the Axoloti, instead sending data back to the PI2, and then driving the Push2 from there.
(this has he the advantage I can have multiple Axolotis, the Push2 and other devices all connected to the PI2 on a USB hub... and the control things centrally on the PI2)
can I ask, why the PI zero? rather than the PI2? the later is a lot more powerful.
if I was planning on using a PI zero, Id probably still get a PI2 initially, develop the software on that, and then transfer it to the PI zero after... it would make development a much more pleasant experience.