Your results
The four pillars
Your overall score says how a face reads. The pillars say why: each groups related measurements and ratings into one score out of 10.
Harmony
How closely your proportions sit inside the ranges linked with balanced faces.
Built from: 13 front ratios (60%) and 14 side-profile measurements (40%).
See its measurements →Angularity
Definition of the jaw, chin and cheekbones.
Built from: The AI's jawline and cheekbone ratings and 4 angularity measurements, blended with 5 detailed assessments once you run them.
See its measurements →Dimorphism
How strongly your features express traits typical for your sex.
Built from: 3 dimorphism measurements plus face width-to-height and jaw width, blended with 10 detailed assessments.
See its measurements →Features
Eyes, nose, skin, symmetry and the details that draw attention.
Built from: The AI's eyes, nose, skin, canthal tilt and symmetry ratings and 5 feature measurements, blended with 41 detailed assessments.
See its measurements →How measurements and assessments combine
Angularity, Dimorphism and Features start from your measurements and the AI feature ratings. When you run a pillar's detailed assessment, it counts for 60% of that pillar and the measurements and ratings for 40%.
| Pillar | Before an assessment | After an assessment |
|---|---|---|
| Harmony | 60% front ratios, 40% side profile | No assessment; always measured |
| Angularity | Ratings + measurements | 60% assessment, 40% ratings + measurements |
| Dimorphism | Measurements | 60% assessment, 40% measurements |
| Features | Ratings + measurements | 60% assessment, 40% ratings + measurements |
One scale for every score
Each ingredient starts on its own scale: a ratio inside its ideal range scores 9 to 10, while the AI's feature ratings centre near 5.5. Before they are combined, each one is mapped onto the same scale as your overall score, so a pillar of 6 means the same as an overall 6: typical. The mapping was fitted on a graded set of test faces (plain, average and model-like). Harmony describes proportions rather than looks, so it keeps its own spread around the middle instead of following attractiveness.