INPUT
One clear front view
Unusable images are rejected instead of receiving a guessed score.
JAWLINE · FREE · ON-DEVICE
A jawline test here returns two things rather than a verdict: the signed turn of the jaw contour at each side, in degrees, and the gonion-to-gonion width as a share of face width. Both are measured from the front, both are reported with the reference value beside them, and neither is part of the 0–100 composite.
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
Three contour vertices per side are used: one up the ramus, one at the jaw corner, one toward the chin. The engine takes the signed turn from the corner-to-ramus vector to the corner-to-chin vector. The canonical face returns −14.69 degrees on the left and +14.69 on the right. The signs are opposite because the contour is mirrored, not because the reference face is lopsided.
02
These are mesh contour vertices, not a clinical gonion located on bone. The vertex indices are published in the engine so the measurement can be reproduced exactly, but a number from this page is not a cephalometric reading and should not be used as one.
03
The second measurement is the horizontal distance between the two jaw-corner vertices, expressed as a share of face width. The canonical face measures 0.775. A larger number means a jaw that is wide relative to the face above it; it does not mean a better or worse jaw.
04
What people usually mean by a sharp jawline is chin projection and the angle of the submental line — both of which are depth. This engine accepts a front view, and chin projection, facial convexity and nasal projection are all listed as not scored for that reason. Reporting a front-view proxy as if it measured projection would be the dishonest option.
05
Neither the gonial angle nor the jaw width feeds the 0–100 score. That number comes from four weighted dimensions — symmetry 35%, facial thirds 25%, facial fifths 25%, golden ratio 15% — and those weights are fixed and published. The jaw measurements sit beside the score as separate readings.
06
The jaw measurements run behind the same quality gates as the score. No face, more than one face, a face too small in frame, a head tilted or turned past threshold, or geometry the engine cannot trust produces a rejection and no numbers at all. A refused measurement is more useful than a confident wrong one.
07
The jaw contour coordinates are computed in your browser and are never transmitted. We do not receive your photo, your landmarks, your angles, or a record that you ran the test, and no account is required.
FAQ
Two readings: the signed gonial angle at each side of the jaw, in degrees, and the gonion-to-gonion jaw width as a share of face width.
The signed turn from the jaw-corner-to-ramus vector to the jaw-corner-to-chin vector, measured on mesh contour vertices. The canonical face returns −14.69 degrees left and +14.69 degrees right.
That is the sign convention, not an asymmetry finding. The contour runs in mirrored directions on the two sides, so a symmetrical jaw produces equal magnitudes with opposite signs.
No. These are face-mesh contour vertices, not bone landmarks. The measurement is reproducible because the indices are published, but it is not a cephalometric measurement.
Not reliably. Sharpness is mostly chin projection and the submental angle, which are depth measurements. This engine measures a front view, and projection is explicitly listed as not scored.
No. The 0–100 composite is built from four fixed weighted dimensions, and the jaw measurements are not among them.
Run the free scan in your browser. You will see the score, fixed weights, measured dimensions, and limits together.