Rate My Face
An AI face rater that shows its work. You get a rating out of ten and, underneath it, the four measurements that produced it — symmetry, two proportions, and face shape.
- No signup
- No upload
- Free
- ~10 seconds
Click to choose a photo, or drop one here
Straight-on, even lighting, at arm's length gives the most comparable reading.
No photo to hand? Try one of these and see what the tool returns.
Runs in your browser. Your photo is never uploaded.
How does a face rater produce a rating?
This one shows its arithmetic. A 478-point face mesh takes the measurements, symmetry is weighted at 40% and three proportion ratios at 20% each, and the 0-100 result is reported on a 1-10 scale. Every component is displayed with the measurement behind it.
- The rating is a weighted average of four measurements, not a model trained on human ratings.
- Because the weights are fixed and published, the same photo always produces the same rating.
- Camera angle and lens distortion move the landmarks and are the main source of variation between photos.
How this AI face rater works
A face mesh places 478 landmark points on your features. From those points the tool computes bilateral symmetry, the ratio of face height to cheekbone width, the ratio of eye spacing to nostril width, and the ratio of nose length to mouth width. Those four numbers are weighted into a rating out of ten.
Nothing is learned, guessed or inferred from a training set of rated faces. It is arithmetic on distances, which is why the same photo always returns the same rating.
Face rating vs. an AI that guesses
Most tools that rate your looks are a language model given a photo and asked for a number. That produces a confident-sounding figure with nothing underneath it, and the same photo will get a different answer on a second run.
Reproducibility is the whole difference between a measurement and a compliment. If a face rating changes when nothing about the face changed, it was never measuring anything.
Rate my looks — what the number leaves out
Geometry is one input, and in daily life it is not the dominant one. Grooming, skin, sleep, posture, expression and how much someone already likes you move perceived attractiveness further than any two-millimetre ratio does.
Skin in particular is worth noting: it has a stronger relationship with rated attractiveness than proportion does in published work, and unlike bone structure it responds to things you control.
Rate my face without uploading a photo
The model runs in your browser. Your photo stays in local memory on your own device, is never transmitted, and is never stored. There is no upload to intercept and no copy for us to hold.
Open your network panel while you run it if you want to confirm — the model file downloads, and nothing containing your image is sent.
How to get a face rating you can actually compare
The rating is deterministic — the same photo returns the same number every time — so any variation you see between runs comes from the photo, not from the model. That makes it worth taking two or three under different conditions and treating the spread as your real range rather than trusting any single result.
Three things move the number most. Head rotation compresses one side of the face and drops the symmetry component. Camera height changes the proportion between the thirds, and a phone held above eye level shortens the lower third measurably. And expression moves the mouth, which feeds the proportion measurements the score is built from. A relaxed face, straight on, at eye level, in even light is not a styling tip here — it is the condition the arithmetic assumes.
What a rating out of ten is actually made of
Four measurements go in: symmetry across matched landmark pairs, face length against width, the eye-to-nose proportion and the nose-to-mouth proportion. Symmetry carries the most weight, and the three proportions are compared against the phi reference rather than against a population average — a convention this site states rather than hides.
That construction is why the number should not be read as a verdict. It summarises four things you can see on the result card, each with its own bar, and the interesting information is which of the four is furthest from the reference rather than what the composite came to. If you only look at the headline figure you are throwing away the part that means something.
Tips for an accurate face rating
- Straight-on photo at arm's length, camera at eye level.
- Even front lighting; side light creates edges the mesh can misread.
- Neutral expression and hair pulled back.
- Skip beauty mode — many phones smooth and reshape by default.
Rate My Face — frequently asked questions
Does the rating change if I smile?
Yes, and by more than you would expect. A smile moves the mouth, which feeds two of the four proportion measurements. The score assumes a relaxed face, and comparing a smiling run to a neutral one is comparing two different inputs.
Is a 7 a good rating?
It is above the midpoint of a scale built on four geometric measurements, and that is all it is. The scale is a convention on top of arithmetic, not a population percentile, so there is no meaningful answer to what a given number is worth.
Is this face rater free?
Yes, with no account. The rating and every measurement behind it are free. Paid credits apply only to the optional simulation feature.
How does the AI rate my face?
It places 478 landmarks, measures symmetry and three proportions, and weights them into a score out of ten. Symmetry carries the most weight. Every component is shown separately.
Is the face rating accurate?
The measurements are exact and reproducible. How well they predict real-world attraction is a loose relationship — expression, grooming and context matter more than geometry does.
Why is my rating lower than I expected?
Usually the camera. A phone held close and low widens the jaw and shortens the midface, which moves two of the four measurements at once. Try a straight-on photo at arm's length.
Can I rate someone else's face?
Technically the tool will process any photo, but do not do it without that person's knowledge. A face is personal data and running someone's photo through a rating tool behind their back is not okay.
References
- 1.Rhodes, G. (2006). The evolutionary psychology of facial beauty. Annual Review of Psychology, 57, 199–226.
- 2.Fink, B., Grammer, K., & Matts, P. J. (2006). Visible skin color distribution plays a role in the perception of age, attractiveness, and health in female faces. Evolution and Human Behavior, 27(6), 433–442.