VPE doesn't use cloud AI or large language models. All intelligence runs locally on your CPU — no GPU needed, no internet required, no latency. Your stream data never leaves your machine.
VPE watches everything that happens on your stream: chat messages, donations, subscriptions, raids, viewer counts, effect firings, and more. Over time, it builds a picture of what your stream looks like and what your audience responds to.
Every decision VPE makes is traceable. You can see exactly why a specific effect fired, what triggered it, and what the AI was thinking. No black boxes, no mysterious choices.