INPUT
One clear front view
Unusable images are rejected instead of receiving a guessed score.
GOLDEN RATIO FACE TEST · FREE · ON-DEVICE
Most golden ratio face tests compare your proportions to 1.618 and score the gap. This one measures the same three ratios and compares them to something else: the values the reference face actually returns. The reference numbers are 1.152, 1.375 and 3.120, and none of them is phi. That is the honest result, and this page is about why.
Free scan · processed on your device · no photo upload · no account
INPUT
One clear front view
Unusable images are rejected instead of receiving a guessed score.
OUTPUT
0–100 and 0–10
Four integer sub-scores, measured values, and reference bands are shown.
PRIVACY
On-device only
Photos and measurements do not leave this browser for the free scan.
01
The golden group is face length to face width, mouth width to nose width, and face width to mouth width. Those three are measurable from a front view with landmarks that correspond to real features, which is why they are the three in the group.
02
Phi describes classical proportion theory stated over anatomical landmarks. What a face-mesh detector returns are mesh vertices, and the two do not line up — the forehead vertex is not a hairline, the face-width vertices are not true zygia. Measured on the canonical face, the three ratios come out at 1.152, 1.375 and 3.120. Scoring you against 1.618 would mean scoring the reference face itself badly, which tells you the ruler is wrong rather than the face.
03
It means distance from those three measured reference values, combined into one deviation and mapped to 0–100. It does not mean proximity to phi, divine proportion, or a beauty standard. A higher number means your three ratios sit closer to one particular average face.
04
This dimension reaches its half-point at a deviation of 0.110 — more than twice the tolerance of symmetry at 0.030, facial thirds at 0.045 and facial fifths at 0.040. It is also the lightest weight in the composite at 15%. Both of those are deliberate: it is the noisiest group, so it is graded the most leniently and counted the least.
05
Each of the three ratios is printed with your measured value and the reference band beside it. If you disagree with a reference value you can see exactly which one you disagree with, instead of arguing with a single unexplained percentage.
06
Symmetry 35%, facial thirds 25%, facial fifths 25%, and golden ratio 15%. The composite is 0.35 × symmetry + 0.25 × thirds + 0.25 × fifths + 0.15 × golden. The weights are fixed in the frozen engine and printed on every page that shows a score.
07
The three ratios are computed in your browser. The photo is not transmitted, the ratios are not transmitted, and no result record is written on our side. There is no account and no payment.
FAQ
No. It compares three measured ratios against the values the engine’s canonical face returns — 1.152, 1.375 and 3.120. Phi describes anatomical landmarks, and a face mesh does not give anatomical landmarks.
Face length to face width, mouth width to nose width, and face width to mouth width. All three are front-view measurements.
It is the noisiest of the four groups on a front view, so it is both the most leniently graded — a half-point at 0.110 deviation — and the least heavily counted.
No. It means three of your measured proportions sit close to one reference average face. The word beauty is not a variable anywhere in the calculation.
Yes. The three ratios are combined into one deviation before scoring, so different combinations can land on the same sub-score. That is why each measured ratio is displayed separately.
Yes, with no sign-up or payment, and the scan runs entirely in your browser.
Run the free scan in your browser. You will see the score, fixed weights, measured dimensions, and limits together.