Who Do You
Mog?

You upload a photo. AI analyzes your facial structure across PSL categories — jawline, canthal tilt, harmony, dimorphism — and compares it against a curated library of 72 reference faces (Henry Cavill, Brad Pitt 1999, Pattinson, Chalamet, Pedro Pascal, etc.). It returns the celebrity whose face most resembles yours (the 'mog match') and the celebrity one PSL tier above whose structure mogs yours.

Curated > exhaustive. The looksmaxxing community references a specific set of ~80 faces — getting matched to a random D-list actor or unknown Korean model is the #1 virality killer in this space. Our 72 are every face that actually shows up in PSL community discussions, era-specific (Pattinson 2008 Twilight vs Pattinson 2022 Batman are separate entries), and PSL-tier-bucketed so the result feels real.

No. Your photo is sent to the AI for analysis, processed in-memory, and deleted immediately after we return your result. We don't store, log, or train any models on your image. The Mogged AI tools have always worked this way.

'You mog [celebrity]' means your face structurally resembles theirs — you share PSL tier and signature traits. 'Mogged by [celebrity]' means the celebrity one tier above you mogs you on specific structural markers. The result shows both — you get the celebrity you're like AND the celebrity who'd beat you in a side-by-side. Same intent as the looksmaxxing community's 'mog' vocabulary.

It's an entertainment classification, not a measurement instrument. The AI is good — modern vision models hit ~90% on identifying public figures and the structural reasoning is solid — but treat the result as a directional snapshot. Run it 2-3 times with different photos for a clearer read. For a full multi-category PSL score, take the photo-based PSL test.

Privacy + legal posture. We don't store or display celebrity images — the AI uses its existing world knowledge to make the match, and the share card shows YOUR photo + the celebrity NAME (not their face). The PSL community actually prefers this format — analysis text is more debatable and screenshot-worthy than side-by-side photo comparisons.

V1 library is currently 72 male-coded references for the male looksmaxxing community. Female and broader libraries are planned for V2. If you're female and run the test, the AI will attempt a match against the male library using structural similarity — you may want to note that result is calibrated for the male reference set.

Three differences: (1) PSL-tier vocabulary — the result places you in HTN / Chadlite / Chad rather than '72% Shawn Mendes'; (2) era-specific celebrity entries — 'Pattinson 2008' and 'Pattinson 2022' are separate, capturing the transformation-arc obsession looksmaxxing audiences care about; (3) honest tier placement — we won't flatter you with a fake Chad match if your structure reads MTN. The community values truth.