They literally proxy your website? I thought they'd cache it... that makes more sense now in your statement that you hit their website with a specially formatted url. Since they pass that through to you you can filter on that.
Also: since you say 4k-5k IPs... any of them from cloud providers? And specific location?
There is also the potential to use it as a watering hole for more sophisticated or subversive measures where they subtly change what you post to promote something you don't actually promote (so at some point they deviate from pure proxy to mitm).
Also: since you say 4k-5k IPs... any of them from cloud providers? And specific location?