Original data on what AI engines actually see when they read a website's images and video. No cherry-picking, methodology shown in the open.
179 images and video combined, 8 live domains, zero cherry-picking. Full per-site breakdown of what's actually machine-readable and what isn't.
We fetched a real homepage the way GPTBot does — no browser, no JS engine. The response was an empty body tag. Here's exactly what came back.
Two comparison tables and a working glossary: Visual GEO vs. traditional image SEO, and accessibility alt text vs. AI-citation alt text.