[fxtentacle]

I'll wait until you have a public API. The trick with reinforcement learning is that I need the feedback from what triggered your system to determine how to modify my image to bypass it.