Looksmaxxing starts with an honest measurement of your own face.

Looksmaxxing means working on your appearance on purpose. The useful version of it starts with a question you cannot answer in a mirror: what is genuinely strong about your face, what is ordinary, and what is holding it back. Canthion answers that by measuring. Twenty-four facial proportions and six skin readings, taken from one photo, in your browser, free. Every number arrives with its margin of error and the population it was compared against, so what comes back is your face rather than a stranger's opinion of it.

Measure my face — free →

What looksmaxxing is, and the two halves of it

The word stretches from a better haircut to jaw surgery, so arguments about whether it works tend to go nowhere. The forums split it in two, and the split is a good one. Softmaxxing is anything that changes how a face reads without touching the bone underneath: skin, hair, brows, facial hair, glasses, body fat, posture, sleep, and how you get photographed. Hardmaxxing is the structural end. Orthodontics, fillers, implants, surgery.

Almost every honest study of how faces are rated puts the first half well ahead of the second. It is also the half you can start on this afternoon, for the price of a razor.

Canthion measures both halves. The proportions that come from bone are measured and graded so you know where you stand: jaw width, midface height, eye spacing, the tilt of the eye. Then the plan is spent on the half that moves.

What the measurement actually covers

One front photo gives Canthion 478 facial landmarks and, from them, 24 proportions used in the aesthetics literature: canthal tilt, facial thirds and fifths, eye spacing, nose proportions, jaw width, facial symmetry, midface ratio, philtrum length, lip proportion, eye openness and more. Each is graded A to F against published norms for your sex and reference group, with the error bar in plain view. Skin is read off the same photograph: redness, texture, spots, dark circles, shine and evenness of tone.

Free gets you the whole harmony read. The paid report adds the score out of 100, your face shape, the masculinity to femininity axis, mock-ups of the recommended looks drawn on your own photo, and a plan.

Add more angles and the report grows. A smiling shot, both three-quarter views, a profile, one with the hair held back, a skin close-up: between them they add jaw taper and definition, the neckline, the ears, the real shape of your hairline, smile width and tooth show, and chin projection traced off the profile edge. Every photograph buys a named read, and the capture screen tells you which one before you take it.

The looksmaxxing terms, measured properly

Most of the forum vocabulary maps onto a real measurement with a real norm behind it. Where it does, Canthion measures it. Where the word names something a photograph cannot see, it says so rather than guessing.

• Canthal tilt, hunter eyes, positive vs negative tilt. The angle from the inner to the outer corner of the eye. Reliably measurable from a front photo, and one of the few eye-area traits that is. How it is measured, and the norms →
• Interpupillary distance, wide-set or close-set eyes. Read as intercanthal distance against palpebral fissure length, which is the ratio the anthropometric literature actually uses. Eye spacing →
• Jaw width, bigonial width, "wide jaw". The width across the jaw angles as a share of the cheekbone width. Measurable head-on; the angle of the jaw is not. Jaw width → · Why a gonial angle from a selfie is a guess →
• FWHR, facial width-to-height ratio. Widely quoted, weakly supported, and sensitive to the exact landmarks chosen. Canthion explains what it can and cannot tell you. FWHR →
• Midface ratio, "long midface", "compact midface". Eye-to-mouth height against cheekbone width, the read behind "long" and "compact" midfaces. Midface ratio → · Face width-to-height →
• Facial thirds and fifths. The classical vertical and horizontal divisions of the face, and how far yours sit from equal. Thirds → · Fifths →
• Orbital vector, "negative orbital vector", under-eye hollows. A profile measurement of where the eye sits relative to the cheek. A front photo does not contain it; the profile shot does, and the report says which. Orbital vector →
• Recessed chin, chin projection. Read from the traced silhouette of a profile photograph, the same way the clinical literature reads it, not from the front-face mesh. Chin projection →
• Golden ratio face, phi mask. A popular test with almost no support in the rating research. Canthion measures the proportions the mask claims to measure, and tells you why the 1.618 itself is not a target. The golden ratio →
• Facial symmetry, asymmetry. Measured landmark by landmark against the midline, with the error bar, because every face is asymmetric and the question is only by how much. Symmetry → · Take the symmetry test →

The full list, with the formula and the source behind each one, is on the methodology page, and the terms are defined in the glossary.

Where the looks actually come from

Bone is a small share of how faces get rated. Skin, grooming, hair, framing and how lean you are make up most of it, and you can change every one of those without a surgeon.

Canthion is built on that. The bone-level proportions are measured honestly and you are told where they sit. The report then spends itself on the part you can move, with every recommendation graded by how good the evidence behind it is. It will not draw you a surgical “after”. A face you cannot obtain is not a plan.

For most faces the levers rank roughly like this, biggest first. The report scores you on every one of them and names the two or three doing the most work against you.

1. 1SkinClarity and evenness of tone. The best-evidenced lever on the list, and the cheapest.
2. 2Body fatDecides how much jaw and cheekbone actually show. Moves faster in the face than anywhere else.
3. 3HairCut, length, and where the hairline sits against the forehead. Fully reversible in a fortnight.
4. 4Brows and facial hairReshape the eye area and the lower face without touching either.
5. 5GlassesRedraw the eye line and the apparent width of the midface. An afternoon at an optician.
6. 6Bone proportionsMeasured and graded so you know where you stand. Nothing on this page moves them.

Bar length is relative weight, not a percentage. The ordering follows the rating literature cited on the methodology page; the exact gaps between rungs vary by study and by face.

How to looksmax from a measurement, step by step

1. Take one good front photo. Even light from the front, no hard shadows, hair off the forehead, neutral expression, camera at eye height and at least an arm's length away so the lens does not widen the nose. A phone in good daylight is fine.
2. Run the free scan. The landmarks and every proportion are computed in your browser in a few seconds. Nothing is uploaded. Start it here.
3. Read the grades, not the score. A single number hides the information. Look for the two or three proportions or skin metrics graded lowest — those are where a change will show.
4. Sort them into movable and fixed. Skin, hair, brows, beard, frames and leanness are movable. Jaw width and eye spacing are not, and no routine will change them.
5. Work the movable column in order of evidence. The complete report grades each recommendation by the strength of the research behind it, so a daily sunscreen and a retinoid, which have decades of evidence, are separated from mewing, which has none.
6. Re-measure the same way in eight weeks. Same light, same distance, same expression. Because the measurement is deterministic, a change in the number is a change in your face, not in the tool.

How this differs from the rating apps

Most looksmaxxing apps give you a number out of ten and some general advice. Upload the same photo twice and the number often changes, which tells you it was never measuring the picture in the first place. Canthion returns the same measurements for the same photo every time, and you can check that against the hash printed on the report. Nothing is sampled and nothing is guessed.

Your photo is also measured on your own device. Landmarks, measurements and the report itself are all computed in the browser. The comparison page goes through the differences properly, and the method behind every value is on the methodology page.

The other difference is what it refuses to do. A photograph cannot measure bone, so you will not get a gonial angle, an E-line or a profile angle out of a selfie here. A tool that quotes those is putting a bony name on a soft-tissue estimate. Canthion reports what a photograph genuinely supports, names the method next to it, and asks for the profile view when the profile view is what the question actually needs.

Common questions

What is looksmaxxing?

Working on your appearance deliberately. Grooming and skincare at the sensible end, surgery at the far end. The version worth your time starts by measuring your own face, so you know where the effort pays, instead of copying a routine built for someone else's.

Is looksmaxxing real, or a scam?

The measurement is real, and so is the effect of skin, grooming and hair. The scams are the scores that change when you upload the same picture twice, and the bone angles quoted off a photograph that cannot measure them. The numbers here reproduce, the method is named beside each one, and the advice is graded by its evidence.

What is the difference between softmaxxing and hardmaxxing?

Softmaxxing changes how a face reads without changing bone: skin, hair, brows, facial hair, glasses, leanness. Hardmaxxing changes the structure itself: orthodontics, fillers, surgery. Softmaxxing accounts for most of how faces are rated and carries almost none of the risk, so that is where Canthion's plan stays.

How do I measure my face for looksmaxxing?

One clear front photo, even light, hair off the forehead, and something that finds the landmarks and compares them against norms. Start the free scan and the read takes a few seconds.

Does mewing work?

No controlled evidence shows tongue posture changing an adult jaw. Canthion grades its recommendations by evidence, and mewing does not make the plan. What does move jaw definition is body-fat level and, for men, the shape of the beard. Both are measurable, and both are in the report.

Does it require surgery?

No. The largest levers are non-surgical, and Canthion's plan stays inside them.

Is the score the same every time?

Yes, for the same photo. Every value is computed the same way from the same 478 landmarks, so one picture gives one answer and you can verify it by hash. Change the photo and some readings will move, because the light, the distance and your expression are part of what you photographed. That is why the advice above is to shoot the same way each time.

Measure my face — free →

One photo. Nothing uploaded. The overview costs nothing and needs no account.

Related: the attractiveness test · what is my face shape · the facial symmetry test · how Canthion compares to Qoves and the rating apps · what a photograph cannot measure.