Am I pretty? Here is what can honestly be measured.

The question assumes prettiness is a property of a face, the way height is a property of a body. It is not. It is a reaction that happens in other people, and reactions vary by culture, by decade, by context and by who is doing the looking.

Researchers have been testing this since the 1970s, usually by having panels rate photographs. Two findings come up again and again. The first is that panels agree with each other more than you would expect - there is a shared component to attractiveness judgement, and symmetry and averageness are part of it. The second is that the shared component is much smaller than the individual one: most of the variation in how any single person rates a face is specific to that rater, not shared with anyone else.

That second finding is the one the face-rating industry never mentions, and it is the one that matters when you are asking about your own face. Even if you knew your exact geometric score, you would not know how any particular person will react to you, because most of their reaction is not about geometry at all.

• Head angle. The biggest single factor, and it has nothing to do with your face. A ten-degree turn compresses one side of the face in the image, and the symmetry pass reads that compression as asymmetry. The panel warns you when your angle is off.
• Expression. Smiling widens the mouth, lifts the cheeks and changes the lower-third measurement all at once. Neutral expressions score higher on this kind of test and are rated lower by humans, which is a good summary of how little the two things overlap.
• Light. Side lighting darkens one half of the face and blurs the landmark positions on that side. Front-on, even light gives the truest measurement.
• Filters. Beauty modes physically move the landmarks - narrowing the jaw, enlarging the eyes. A filtered photo scores higher and means nothing.
• Genuine asymmetry. Everyone has some. It is the last thing on this list because it is the smallest contributor to a surprising score, not the first.

Q:Am I pretty? Can a photo test answer that?

Not the way you mean it. A photo test measures geometry: symmetry across the centre line of your face, how evenly your face divides into thirds and fifths, and how your proportions compare to classical targets. It is genuinely measuring something, and it measures it precisely. But 'pretty' is a judgement other people make, and the research on what drives that judgement puts expression, skin, grooming, styling, colouring, movement and familiarity well ahead of bone geometry. A test can hand you numbers. It cannot hand you the answer.

Q:Is this am I pretty test accurate?

The measurement is accurate and repeatable - the same photo gives the same numbers to the decimal, every time. What is not accurate is treating that number as a rating of you. Most tools in this category blur the two together deliberately, because a confident-sounding verdict is more shareable than an honest measurement. This one keeps them apart.

Q:Do I have to upload my photo?

Not for the test. The face-landmark model downloads to your browser, your browser reads the photo locally, and the image never travels anywhere; closing the tab is all the deletion that is needed. The only thing on the site that ever sends a photo is the optional paid report, and only if you choose it after seeing your result.

Q:Is the am I pretty test free?

Yes. No account, no email, no credits, and the result is never held back. The optional written report offered under it is the one paid thing on the site.

Q:Why does the test not just tell me yes or no?

Because 'yes' and 'no' are both dishonest. The tool has measured symmetry and proportion in one photograph. Turning that into a yes or a no would require it to know how people respond to faces, which it does not, and to know how you look in motion, in person, on a good day, which no photo carries. Every site that gives you the confident yes or no is making it up.

Q:Am I pretty enough to be a model?

Modelling agencies do not select on facial symmetry scores. They select on height, measurements, bone structure that photographs a particular way, an unusual look rather than an average one, and whatever a specific client needs that season. Faces that measure as highly symmetric are often described as pleasant rather than striking, and 'striking' is what most editorial casting looks for. This test cannot answer the question and neither can any other test - only agencies can.

Q:My score was lower than I expected. What should I do about it?

Check the head angle warning first, because a turned or tilted photo drops the symmetry score on its own and that is the single most common cause of a surprising result. Beyond that: nothing. This site does not sell a fix and will not suggest one. Facial asymmetry is close to universal - almost nobody has a face that mirrors itself, and the small differences that lower the score here are the same ones people describe as character in a face they like.

Q:How is this different from PrettyScale or an AI face rater?

Three things. The method is published rather than hidden, so you can read exactly how the number is produced. Nothing is uploaded, so there is no photo of your face sitting on anyone's server. And the score itself has no job to do other than be accurate - the free result is never held back to sell the report, there is no app to install, and no reason for the number to be tuned to make you feel a particular way.

• How pretty am I? The main test
• Pretty or ugly? Why this site refuses the binary
• Am I pretty quiz: measurement instead of multiple choice
• The 1-10 pretty scale explained