overall mismatch: 1.43%
run at 2026-05-05T23:45:39.513Z
Pixel-diff is sensitive to lazy-loading timing. On sites with IntersectionObserver-driven image loads, the mirror often renders MORE content than the original (assets are local, no network latency) — and the mismatch reflects that gap, not mirror infidelity. For static sites the metric is meaningful (Framer-class typically <2%); for heavily progressive sites expect 10–25% even with a perfect mirror.
