Regarding the "How do you not hear yourself and just listen for the server sound?", I think you figured it out already. But for those who don't know what it means:
Consider Jamulus or Jamkazam your DAW.
There are 2 types of audio interfaces. Those with direct monitoring and those without.
When you record your Kemper tones into your DAW, you are either listening to your direct output or you have monitoring on in your DAW. Monitoring through your DAW introduces some delay. Which is normal as the signal goes through several stages of converters to eventually get into your ears.
If your audio interface has direct monitoring, then in all cases you want to use that. It simply means the signal goes from your Kemper to your audio interface and that's when your hear it. It skips the computer altogether.
You can perfectly fine record your guitar this way. In real time. Because when the DAW plays back the track, in reality it has a slight delay, but it doesn't matter to you. You can play along or record the 2nd guitar track and the DAW makes sure it's in sync with the recorded part.
If you don't have a direct monitoring option, then you have to adjust to how fast your audio interface and computer is. You have to tell your DAW to enable "input monitoring". Which translates to: I send info to my Kemper, my Kemper sends it to my audio interface > my audio interface can't let it hear me directly so it sends the signal to the DAW. The DAW has to process it and send it back to my audio interface.
So this typically introduces latency. And depending on your interface, anywhere between 2 and say, 12 ms. And this is if you have your settings configured for live use. For mixing you want to ease these settings to give your CPU some breath. But that's another topic.
Anyway, the average audio interface should be able to handle 8 ms, without getting sound artifacts. Usually, the lower you go, the more your computer has to work, the more your sound goes bad.
Anyway, the reason I explained this is, is because this also explains why you want to go through the Jamkazam or Jamulus server before hearing your own sound. You'll get in sync with the rest of the band!
And how to achieve this is yet another big post. Which I'm not gonna do right now.
The most easy way is to tell the Jam software to insert A and output B. Where A and B are the inputs and outputs of your audio interface.
I personally like to use several instruments and mics. So I go through the Reaper software. I add as many channels as I want. I go into Jamkazam, with 10 instruments if I have to. Then Jamkazam sends the signal back to Reaper, where I can record it. And I listen to the return signal, through my headphones. And play along. On my guitar, electronic drums or whatever.
And because I listen to the return signal, I'm pretty much perfectly in sync with the people I'm playing with!