Real pixel-level checks on your actual image — our own heuristic score, not a claimed AI attention model.
Paste a YouTube video URL (we'll pull its real thumbnail) or upload your own image. Every number comes from actually reading the image's pixels in your browser.
Resolution, contrast, and complexity are genuinely measurable and genuinely useful — but none of them predict whether a specific audience will actually click. That still takes human judgment.
Real pixel data from your image: resolution, aspect ratio, contrast (brightness standard deviation), and a visual-complexity estimate (edge density) — all computed live in your browser, not by an AI model.
Treat each check as a concrete, fixable property, not a verdict on whether the thumbnail will work.
All analysis happens in your browser via the Canvas API — your image is never uploaded to a server for this specific check.
No, and we won't claim it is. A real attention-prediction model needs training data and infrastructure we don't have — faking that score would just be a guess dressed up as AI. What you get instead is our own heuristic score built from properties we can actually measure by reading the image's real pixels: resolution, contrast, brightness, and visual complexity.
An edge-density estimate across the image — lots of sharp brightness changes usually means a busy, cluttered composition; fewer, concentrated edges usually means a cleaner single focal point. It's a real signal, but it's a proxy, not a judgment of whether your thumbnail is "good."
We build, edit, and grow fully automated YouTube channels end-to-end. Strategy, production, SEO, and monetization. You own the asset.