Skip to content
The Face Shape Detector
Policy

Disclaimer

The Face Shape Detector tools are for entertainment and information only. Read what our face, style and attractiveness results can and cannot tell you, and why

Last updated

Effective date: July 22, 2026

Last updated: July 22, 2026

A disclaimer states the boundary of what a publisher claims. The Face Shape Detector runs 14 free face and style analysis tools plus articles on facial proportion, color analysis and style frameworks. This page marks where the claims stop.

Everything published here is entertainment, not assessment. The photo tools place 468 landmark points with Google MediaPipe, entirely inside your browser, and compute ratios from those coordinates. No photograph is transmitted to, received by, or stored on our servers. Ratios are arithmetic, and the arithmetic is exact. What it cannot do is diagnose, assess, or tell you anything about your health, your worth or your future.

The categories those ratios feed are style vocabularies rather than scientific taxonomies. Their sources appear further down: beauty writing of the last century, David Kibbe's 1987 book, seasonal color theory, and W. H. Sheldon's discredited somatotypes.

Below: that framing in full, the absence of medical or professional advice, seven factors that degrade landmark accuracy, why the classification systems are frameworks rather than science, a blunt section on attractiveness scoring and age estimation, outbound links, advertising and affiliate disclosure, availability limits, testimonials, and how to report something wrong.

This is a policy document, not legal advice. Read it before acting on anything the site tells you. If you read one section, read the one on attractiveness scoring.

Everything here is for entertainment and general information

Entertainment is the honest description of every tool here: face shape detector, eye shape detector, nose shape detector, lip shape analyzer, face symmetry test, golden ratio face, color analysis quiz, skin tone analyzer, body shape calculator, body type calculator, Kibbe body types, attractiveness test, celebrity look-alike finder and age estimator. They exist to be interesting, and occasionally useful when you're choosing glasses, a haircut or a palette. Two report as unavailable, the look-alike finder and age estimation, because each needs a trained model that landmark geometry cannot replace.

None of it is a diagnosis, an assessment, a measurement of record, or a professional opinion. Software generates every result with no human review. Your medical history, goals, budget and context are unknown to us, and the code makes no attempt to infer them.

Treat a result the way you'd treat a magazine quiz that runs on geometry: a starting point for your own judgment, never a conclusion. And never a substitute for a professional.

No medical or professional advice

Medical advice is not what this is. Nothing here is dermatological, dental, orthodontic, surgical, psychological, psychiatric, nutritional, fitness, or any other professional guidance, and nothing on the site creates a clinician-patient, therapist-client or adviser-client relationship.

If you have a health concern of any sort, including concerns about your skin, weight, body image, eating, mood or appearance, consult a qualified professional who can examine you and take a history. Never disregard professional advice, and never delay seeking it, because of something you read here.

Specific things this software cannot do, in plain terms:

  • It cannot detect or rule out any skin condition, mole, lesion or cancer. The skin tone analyzer estimates a color value from pixels and holds no clinical capability.
  • It cannot read facial asymmetry as a clinical sign. Asymmetry appearing suddenly, or alongside weakness, numbness, drooping or pain, needs urgent medical attention. That is an emergency, not a symmetry score.
  • It cannot assess body composition, metabolic health or nutritional status. The body shape and body type calculators work from ratios you type in.
  • It cannot assess mental health. If a result distresses you, or feeds a preoccupation with your appearance, close the tool and speak to a doctor or a mental health professional.

In immediate danger or crisis, contact your local emergency number or a crisis line now, not us. Urgent help is beyond us, and the contact form goes unmonitored outside business hours. Ordinary results have limits too.

Accuracy limits of the analysis

Accuracy depends on the photograph. The photo tools use Google MediaPipe to place 468 landmark points and compute ratios between them, in your browser, with no image transmitted anywhere. The technique is well established. It also has predictable limits.

FactorEffect on the resultWhat to do about it
Head angle and tiltA few degrees of yaw or pitch shifts width-to-height ratios enough to move a borderline face between two categoriesFace the camera square on, eyes level
Lighting and shadowStrong side light darkens one jaw edge and moves the perceived outlineUse even, front-facing light; avoid overhead spots
ExpressionA smile widens the cheek and lower face; raised brows change the forehead measurementKeep a neutral, relaxed expression
Hair, fringe and headwearHair over the forehead or jaw hides the outline the model readsPull hair back where you can; expect wider error bands if you cannot
GlassesFrames bias eye landmark placement and the eye shape resultRemove glasses for eye shape; keep them on if you are choosing frames
Camera and lensPhone selfie lenses at close range enlarge central features relative to the outlineHold the camera at arm's length or further
Image resolutionLow-resolution or heavily compressed images degrade landmark precisionUse a photo at least 600 pixels across the face

Percentage confidences get published for this reason. If your top result reads 41 percent and the next 37 percent, the honest reading is that you sit between two categories, not that the software is 41 percent sure of one. The categories carry a deeper problem still.

The classification systems are style frameworks, not science

Frameworks, not taxonomies. Even with perfect geometry, the categories these numbers feed have no scientific standing.

Face shape names (oval, round, square, heart, diamond, oblong, triangle) come from twentieth-century beauty and styling literature. No agreed clinical definition exists, and no standard boundary separates one from the next. Sources draw the lines differently, which is why two websites hand you two answers from one photograph and both stay internally consistent.

Kibbe body types trace to a 1987 style book. Seasonal analysis descends from mid-century color theory popularized in the 1980s. Somatotypes, the ectomorph, mesomorph and endomorph triad, originate in Sheldon's 1940s constitutional psychology, whose claims linking build to temperament have been thoroughly discredited and appear nowhere here. We borrow the vocabulary because it helps in discussing proportion and styling. We present none of it as biology.

The golden ratio material sits in the same position. The 1.618 proportion is real and elegant mathematics with a long history in art and design. The claim that faces closer to it are objectively more beautiful is unestablished, and deciding where a face begins and ends is subjective enough to yield whatever number you look for. Which brings us to the section that matters most.

The attractiveness test and the age estimator

Read this carefully.

The attractiveness test does not measure beauty. It computes how close a set of measured facial ratios sits to a statistical average drawn from population data, then expresses that distance as a score. That is all it does. Averageness and symmetry correlate weakly with rated attractiveness in some published studies, in some populations, under laboratory conditions. That sits a long way from how you look to another human being.

A score from this tool is not a measure of your beauty, worth, health, desirability, prospects, or anything else about you as a person. It cannot be. It holds nothing about your voice, your expression in motion, your warmth, your style, your humor or your presence, which is most of what people respond to.

Beauty standards are not fixed. They differ profoundly between cultures, and they have shifted inside every culture across decades and centuries. Features prized in one place and time were unremarkable in another. Any single numeric score is at best a snapshot of one narrow statistical construct built from one dataset.

Age estimation is currently listed as unavailable for the reason given above, and the caution applies wherever such a tool runs. Estimators compare an image against patterns in a training distribution and return the age those patterns resemble. They are frequently wrong by five years or more, and systematically less accurate for some ages, ethnicities and lighting conditions than others, because the underlying data is unevenly distributed. A high number means the photograph resembled photographs of older faces. Nothing more.

A direct note to younger readers and to anyone who is struggling. If you are a teenager, or if you already spend a lot of time worrying about how you look, please don't take a number from this site as a judgment of yourself. It is not one. It cannot be one. Software that has never met you cannot tell you what you are worth. If a score here has made you feel worse about yourself, close the tab and talk to someone you trust, or to a doctor. That is a far better use of the next ten minutes than another photograph.

Links in our articles point to research papers, retailers, style resources and reference material. Those sites are outside our control, and responsibility for their content, accuracy, availability, privacy practices and security rests with them.

A link is a pointer, never a recommendation. Follow one and you leave The Face Shape Detector, at which point that site's policies apply. Check them before entering personal information or making a purchase. Advertisements follow a similar principle.

Advertising and affiliate disclosure

Advertising pays for everything here. Ads are served through Google AdSense, selected by Google rather than vetted by us, and their appearance beside a piece of content endorses neither the advertiser, the product, nor any claim it makes. If an ad looks like it breaches policy, report it and we'll pass it to Google.

Some article links may be affiliate links, meaning we could earn a small commission if you buy through them at no extra cost. Where that applies, the page discloses it. Commercial relationships never determine an editorial conclusion, and our editorial policy explains how that separation holds. Revenue keeps the tools free, though free guarantees no availability.

Availability and technical performance

Availability comes on an "as is" and "as available" basis. Uninterrupted access, freedom from errors, and correct operation on a particular device, browser or operating system are none of them guaranteed.

Because analysis runs on your own hardware, performance follows your device. Older phones may be slow or fail to load the landmark model. Browser extensions, privacy modes and corporate network filters can block the model file and stop a tool starting. Features may change or disappear without notice.

To the fullest extent permitted by law, we exclude liability for any loss arising from your use of, or inability to use, this site, including any decision about your appearance, your health, a purchase or a professional service made on the strength of a result. Nothing here excludes liability that cannot lawfully be excluded, including liability for death or personal injury caused by negligence or for fraud.

Editor note: replace this paragraph with the registered legal entity name, jurisdiction and governing law clause supplied by your legal adviser, so the limitation above is enforceable in your operating jurisdiction.

Testimonials and results

Testimonials reflect the experience of whoever gave them. Individual experiences vary, and nothing in a quoted comment promises you will find a result accurate, useful or flattering.

Outcomes carry no claim from us. A face shape result guarantees nothing about whether a recommended haircut suits you, and a color season guarantees nothing about how a palette looks under your local light. Use your own eyes and, where it matters, a professional's. Where a result looks plainly wrong, tell us.

Reporting an inaccuracy

Reports are welcome. If a tool returned a result you believe is wrong, or an article contains a factual error, email [email protected] or use the contact form and include:

  1. The exact page URL.
  2. What you expected, and what came back instead.
  3. Your browser, browser version, device and operating system.
  4. The photograph conditions (lighting, angle, glasses, hair) where the report concerns a photo tool.

Do not send us a photograph. Images cannot reach us, our systems are not built to accept them, and we don't want them by email. A description is enough to reproduce and investigate.

Replies usually arrive within two to three business days, and confirmed article errors get corrected with a dated note rather than a silent edit.

The boundary in short: everything on The Face Shape Detector is entertainment and general information, generated by software with no human review, and none of it is medical, dental, psychological or any other professional advice. The photo tools place 468 landmark points in your browser and compute exact ratios, but head angle, lighting, expression, hair, glasses, lens and resolution all move the answer, which is why confidence percentages appear instead of a single verdict.

The categories those ratios feed are style vocabularies borrowed from beauty writing, a 1987 book, seasonal color theory and Sheldon's discredited somatotypes. Attractiveness scoring measures distance from a statistical average, nothing about your worth. Outbound links are pointers, not recommendations. Google AdSense selects the ads. Availability is "as is". Testimonials promise nothing.

Found something wrong? Email [email protected] with the page URL and the conditions. Keep the photograph.

Questions about this policy?

If anything here is unclear, or you want to know what we hold about you, write to us and we will answer.