Face Rating Test

A face rating test uses AI facial landmark detection to measure specific geometric properties of your face — like symmetry and proportion — against established aesthetic reference ranges. This tool reports each measured signal on its own, rather than combining them into a single number.

The Attractiveness Test (/attractiveness-test) scores five facial proportion ratios — eye spacing, nose breadth, mouth width, face ratio, and facial thirds — and combines them into a single 0-100 score. This Face Rating Test uses a different set of five signals — symmetry, golden-ratio proximity, canthal tilt, jaw definition, and eye harmony — and deliberately does not combine them into one number, showing each signal's own modeled percentile instead. They measure different things and will often give different results for the same photo.

No. This is an illustrative, entertainment-focused tool. It measures real geometric properties of your face using real landmark detection, but the "ideal" reference ranges (golden ratio, canthal tilt range, etc.) are drawn from popular aesthetic literature, not a rigorous scientific consensus on attractiveness — which does not exist. Each signal's percentile is modeled against an assumed population curve, not real survey data.

No. All processing happens entirely in your browser using Google's MediaPipe Face Landmarker. Your photo never leaves your device and is not uploaded to any server.

Canthal tilt is the angle of the line running from the inner corner of your eye to the outer corner. A positive (upturned) tilt, often cited as +2° to +8°, is frequently referenced in beauty and cosmetic surgery discussions as aesthetically favored — though preferences vary widely across cultures and individuals.

The golden ratio (phi, approximately 1.618) is a mathematical proportion that recurs in classical art and architecture. Some facial-aesthetics frameworks claim that facial proportions closer to phi — like face height to width — are perceived as more harmonious, though this claim is debated in the scientific literature.

Averaging five unrelated geometric measurements into a single number implies a precision and meaning the underlying math doesn't actually have — it would read as an authoritative verdict when it's really five independent, debatable estimates. Showing each signal separately is more honest about what's actually being measured.

Yes, significantly. Head angle, expression, lighting, camera distance, and even image compression can all shift the detected landmark positions and therefore every signal in this test. For a consistent result, use a similar front-facing, neutral-expression photo each time.

Each signal's percentile is an illustrative estimate of how that specific measurement compares to a modeled population curve (assumed to be roughly centered around 62/100, with typical variation). It is not based on real survey data of other users' scores — treat it as a fun indicator, not a statistic.

The landmark detection itself works across face shapes and ethnicities. However, the "ideal" reference ranges used for scoring (golden ratio, canthal tilt, etc.) originate mostly from Western aesthetic traditions and may not reflect the full diversity of features and beauty standards worldwide — keep that context in mind when reading your result.

This tool covers the same general idea as what online communities call a "PSL score test" — rating facial attractiveness from a photo — but it is not affiliated with any specific forum or rating system, and it does not use PSL's informal scoring conventions or produce a single combined number. It reports five independent, geometry-based signals (symmetry, golden ratio, canthal tilt, jaw definition, and eye harmony), for entertainment purposes only.