I'm not familiar with Tonedexter's input impedance. But the third channel taking input from my Gage Lifeline pickup enters a Noble tube preamp with 10MegOhm input impedance and its XLR output impedance of 420ohm (
@30hz) goes straight to a Zoom F8 solid state field recorder or Focusrite interface.
The "regular" input impedance of the TD is 1MegOhm; however, if you use a TRS cable to the input, the smart jack will change the input impedance to 10MegOhm. So there's that.
The input impedance of the first thing you plug the pickup into is the important thing, the impedance of all of the other gear should be relatively inconsequential. So, if you went ahead with this experiment, using the Noble pre as your front end, recorded the "training exercises" then reamped the mic signal against the pickup signal to train the ToneDexter, you'd probably want to use a TRS cable to then use the TD's ToneMap for live performance moving forward, as that would be the closest conditions to those you trained the unit in.
Note that any tone coloration inherent in the Noble pre would create some slight variances in the ToneMap, since you wouldn't be using it (presumably) for live performance. You know, things that make it sound "Tube-y" or "Vintage."
Altogether, it presents as an interesting experiment. If you should proceed with it, I look forward to hearing how the results pan out.
Personally, I do think that it would be much easier to just skip the recording step; get "your tone" with the two mics into a mixing board, and then send that signal - live - to the ToneDexter. While I know you could more carefully "fine tune" the blend using a recording, it seems to me that the other variables will affect that.
Meaning, say that the live mix gets you 95% of the way to the perfect tone, while recording allows you to dial it in to 100%; but the Noble pre, slightly different impedances, and other added gear factors create a skew of 10% from the intended sound once you finally get the ToneMap created. The live version still wins, yeah? And the intended subtleties will likely get "lost in the sauce" within the context of a band performance in a live venue.