Your next step
Score your own face free
Upload one selfie, get scored on 17 facial-geometry metrics + your validated Impression Percentile in ~30 seconds. No signup. Full written report $14.99.
Score my face free →Upload a selfie
Get scored, free
Unlock the report
The only validated face test in the category
Most "am I ugly" tools are golden-ratio calculators that never check their number against how real people actually rate faces. RealSmile is the only one that does. Your Face Score is a validated attractiveness percentile — it agrees with averaged human ratings at r ≈ 0.8, cross-validated against 5,500 human-rated faces in the SCUT-FBP5500 academic benchmark (60 ratings per face). See the full validation →
The honest answer to "am I ugly?"
Short answer: almost certainly not — not in the way you fear. "Ugly" isn't a fixed trait; it's a continuous spectrum. Most people who ask "am I ugly?" or "how ugly am I?" turn out to have one or two specific, usually fixable features scoring below average — not an overall deficiency. The honest way to know is to measure your face objectively, not to judge it in a distorted mirror. Last updated June 2026.
"Ugly" is not a binary. Research in facial perception consistently shows that attractiveness is a continuous spectrum shaped by dozens of factors — many of which change. The way most people evaluate themselves is also systematically distorted: mirror reversal, phone camera lens distortion, bad lighting, and the familiarity bias (we're used to our own face and see flaws others don't notice) all make self-assessment unreliable.
What AI face analysis actually does is measure specific, objective features: bilateral symmetry (the facial symmetry test measures this one on its own), canthal tilt angle, facial width-to-height ratio, jawline definition (see the jawline test for a free jawline rating), and facial thirds proportion (the free facial thirds test measures your upper, middle, and lower third split on its own). These are the same metrics used in aesthetic medicine and published attractiveness research. A low score on any of these tells you something specific and actionable — not a verdict on your worth. That's also what separates this from a generic "ugly tester" that spits out a yes-or-no label: you get measurements of your photo, not a judgment of you as a person.
The most common finding: people who feel "ugly" typically have 1-2 specific features that score below average, not an overall deficiency. Identifying those specific features is far more useful than a general "ugly or not" judgment. If you want the inverse framing, run the am i pretty test instead — same engine, different question. Optimizing for work rather than dating? The LinkedIn / Executive Pro Audit scores the same 17 metrics weighted for competence and recruiter confidence.
What your result actually means
Your scan gives you two different things, and reading them correctly is the whole point:
Your Face Score — the number to trust
A validated attractiveness percentile: it tells you where a still photo of your face lands against a large population. This is the one output built to answer "am I ugly," because it is cross-validated at r ≈ 0.8 against how real people rate faces. A percentile is a position, not a verdict on your worth — and most people who searched "am I ugly" land closer to the middle than they expect.
The 17-metric map — measurements, not a rank
Symmetry, canthal tilt, jawline, facial thirds and 13 more: a map of how each feature measures against population norms. It is not a ranking. We tested blending these into a single attractiveness number and it did not predict how people rate faces (r ≈ −0.33), so we do not present it as a score. Use it to find the one or two features scoring below median — the specific, usually fixable things worth your attention.
Put together, the honest read for almost everyone who lands on this page is: you are not "ugly." You have a Face Score that is probably more average than your inner critic claims, plus one or two specific features you have been over-weighting. That is a far more useful — and more hopeful — answer than a single yes-or-no label. If you want the encouraging framing of the same data, the am i pretty test runs the identical engine on the opposite question.
Is this "am I ugly" test accurate?
This is the question that actually matters, and it deserves a real answer rather than a marketing one. Most face-rating tools are golden-ratio calculators: they compare your proportions to a fixed template and never check whether that template matches how real people rate real faces. On the standard academic benchmark, that golden-ratio approach reaches only about r ≈ 0.55-0.65 correlation with human ratings — which is why a "beauty calculator" can hand you a harsh number that no actual person would agree with.
RealSmile's Face Score is measured differently. We ran our production pipeline against SCUT-FBP5500 — 5,500 face photos, each independently rated for attractiveness by 60 people — using 5-fold cross-validation, so every number is scored on faces the model never saw while fitting. The Face Score reaches a cross-validated r ≈ 0.8 against the averaged human ratings (n = 1,566 held-out faces, consistent across male and female subsets). For reference, the best published deep-learning models on this benchmark reach r ≈ 0.85-0.90.
We hold ourselves to the same test. Our original 17-metric geometry composite failed it — no positive correlation with human ratings (r ≈ −0.33) — so we published that finding and stopped presenting the geometry as an attractiveness score. To our knowledge, no other consumer face-rating tool publishes any validation of its scoring against human ratings at all. If you'd rather see the same validated score on the familiar scale, the rate my face out of 10 page maps your percentile to a 1–10 rating.
Methodology & editorial
Every scan uses 68-point iBUG 300-W facial-landmark detection to place the geometry, then the Face Score model estimates the percentile. Photos are processed in memory and never stored (on desktop you can verify this in your browser's Network tab). This page is maintained by Randy, founder of RealSmile; the full study list and validation write-up are public on the research base, and the citation chain — Rhodes 2006 on averageness, Willis & Todorov 2006 on 100-ms first impressions, the lighting studies — is checkable there.
That is the honest answer on accuracy. The fastest way to stop guessing whether you're ugly is to measure your own face — free, private, about a minute.
Free · No signup · Private — never stored
Why you look worse in photos than in real life
Mirror reversal
You've seen your face in a mirror thousands of times — always flipped. Photos show the non-reversed version, which looks subtly "wrong" to you because it's unfamiliar. Other people only ever see your non-flipped face and have no reference for the reversed version. The "ugly in photos" feeling is largely familiarity bias.
Wide-angle lens distortion
Most phone selfies are taken at 20-30cm from the face using a wide-angle lens (~28mm equivalent). At this distance, perspective distortion makes noses look 30% larger and distorts facial proportions significantly. Professional portraits are taken at 85-135mm at 1.5-2m distance. Take photos at arm's length (60-80cm) to eliminate most distortion.
Bad lighting
Overhead fluorescent lighting, bathroom ceiling lights, and direct flash all cast unflattering shadows that reduce apparent facial symmetry, deepen under-eye circles, and make skin texture more visible. The same face under golden hour outdoor light or window light scores 1-2 standard deviations higher on attractiveness in research. Lighting matters more than most structural features.
Unnatural expression
Posed smiles for photos look different from natural smiles — the timing and muscle activation patterns are off. A genuine Duchenne smile (that reaches the eyes) scores significantly higher on warmth and attractiveness than a camera-triggered pose. The solution is thinking of something genuinely funny or happy in the moment before the photo.
"Am I ugly, or is it my photo?" Usually the photo carries more than you think. Lighting direction alone can shift how symmetric a face looks by up to about 23% independent of the underlying bone structure (Zaidel & Cohen 2005), and first-impression research finds lighting and pose move the read more than the structure itself does (Todorov & Porter 2014). Before concluding anything about your actual face, re-scan a well-lit, arm's-length photo. Sources on our research base.
What AI actually measures when analyzing your face
Unlike "hot or not" voting apps that aggregate subjective opinions, AI face analysis measures specific geometric properties that correlate with attractiveness ratings across populations. Here's what RealSmile's looksmaxxing test measures:
Facial Symmetry
Bilateral balance between left and right halves. Asymmetry above 8% correlates with lower attractiveness ratings.
Canthal Tilt
Angle between inner and outer eye corners. Positive tilt (+3° to +8°) is consistently preferred.
FWHR
Facial width-to-height ratio. Linked to perceived dominance and masculinity in men.
Jawline Angle
Gonial angle measurement. Ideal range 120-135°. Most people measure 130-145°.
Facial Thirds
Equal thirds from hairline to brow, brow to nose base, nose base to chin. Balanced thirds = harmony.
Midface Ratio
Relationship between nose width and inter-eye distance. Narrow midface is generally preferred.
Hunter Eye Index
Composite of canthal tilt, eyelid hooding, and scleral show. The "hunter eye" aesthetic.
Eye Spacing
Distance between eyes relative to face width. Ideal ratio is 1:1:1 (eye:gap:eye).
What you can actually change
Highly changeable
- ✓Body fat % — reveals bone structure, sharpens jawline
- ✓Skin quality — texture, tone, under-eye circles
- ✓Photo technique — lighting, distance, angle
- ✓Expression — genuine vs. posed smile
- ✓Posture — affects jawline angle dramatically
- ✓Grooming — hair, beard line, eyebrow shape
Fixed (without surgery)
- ·Bone structure — orbital shape, cheekbones, jaw
- ·Canthal tilt — determined by orbital rim anatomy
- ·Facial thirds ratios — set by skull development
- ·Eye size and shape — genetic, orbital structure
- ·Nose bone shape — cartilage and bone
Key insight: The changeable factors — body fat, skin, photo quality, expression, posture — collectively account for a larger portion of perceived attractiveness than bone structure in most cases. This is why identical twins raised differently can look noticeably different in attractiveness ratings.
Worried the low number is just the photo, not your face? The free am I photogenic test measures exactly that gap — how much of your score is camera-dependent versus structural.
Most of the "highly changeable" column responds to inexpensive routine changes — skin, definition, grooming. Our best looksmaxxing products guide covers what is actually worth buying for each, and what to skip.
What this test does NOT measure
Honesty matters more than hype. AI face analysis measures static photo geometry — that's a real but narrow signal. Below is exactly what it cannot see, so you can read your score without overweighting it.
Personality & charisma
Humor, warmth, eye contact, voice tone, and the way you carry a conversation are routinely ranked above static facial features in real-world attraction research (Rhodes 2006, PMID 16318594). A photo cannot capture any of this.
Movement & expression dynamics
How your face moves when you laugh, listen, or speak is a separate channel from static geometry. Two faces with identical still-frame scores can land very differently on video. The test only sees one frozen frame.
Body, height, fitness, posture
Roughly the entire signal below the chin is excluded. Body composition, posture, gait, and overall fitness contribute meaningfully to in-person impression and are invisible to a face crop.
Style, grooming, fit of clothing
Hair cut and color, beard line, eyebrow shape, glasses, and clothing fit can move perceived attractiveness by a wide margin (Geniole 2015 on facial structure context, PMID 26181579). The test sees raw face only — not your styled version.
Cultural & individual preference
The model is trained on aggregate ratings. Specific cultures, subcultures, and individual partners weight features very differently. A score that reads "average" globally can be highly preferred in your actual dating market.
Status, context, social proof
Where someone meets you, who you're with, and what you do shift perceived attractiveness substantially. None of this enters a single-photo analysis.
Skin under makeup, filters, or compression
Heavy filters, beauty modes, and aggressive JPEG compression smooth real signal out of the photo. The model reads what the file shows — not your true skin or symmetry.
Read the score as "here is what a still photo of my face geometry suggests" — not "here is how attractive I am." The list above is most of the gap between those two sentences. The same one-photo caveat applies to perceived age — the free how old do I look test reads an age estimate from the same landmarks, and it swings photo-to-photo for the same reasons.
Find your specific low-scoring metrics
Get your actual face analysis
17 metrics measured from a single selfie. Free scan shows your validated Face Score percentile and your strongest metric. Full report with all 17 measurements, your weakest metric with fixes, and a personalized improvement plan for $14.99.
The picture test is free — no signup
The question behind the question
"Am I ugly" is almost never the actual question. It is a stand-in for something specific that a literal score cannot answer. Use the ladder below to translate the search into a question your face score can actually settle.
"Am I ugly?"
What you are actually asking
Why do I feel invisible at parties / on dating apps?
What to do instead
The visibility issue is rarely your overall geometry — it is a specific photo or a specific approach pattern. Score your photos individually and look at trait drift across them, not a single overall number.
"Am I a 3 or a 7?"
What you are actually asking
Where do I land in the population of faces I am competing with?
What to do instead
Composite score and 1-10 self-rating are noisy. Population percentile per trait is the actual signal — it tells you which traits are below median and worth attention.
"Why am I always single?"
What you are actually asking
Is the photo carrying my profile, or is the bio / approach strategy the bottleneck?
What to do instead
Run a face score. If you land above the 50th percentile, the photo is not the bottleneck — switch attention to bio, opener, and platform fit. Most singleness has a non-photo cause.
"What can I change?"
What you are actually asking
Which 1-2 traits are below median AND have a non-surgical lever?
What to do instead
Score yourself. Identify lowest-percentile traits. Cross-reference against the non-surgical intervention list (grooming, posture, weight, lighting, eyewear). Surgery is the last resort, not the first.
"Am I uglier than I think?"
What you are actually asking
Is the gap between my self-image and external read actually a gap?
What to do instead
The dysmorphia-vs-reality gap is real both ways — you can over-rate or under-rate yourself. A neutral score plus 5-10 outside readings calibrates the gap and tells you whether your self-image is the actual problem.
The free score above answers the literal question. The ladder above is for the question underneath — which is usually the one that actually changes your week.
Frequently asked questions
Am I ugly?
Almost certainly not in the way you fear. Attractiveness is a continuous spectrum, and self-perception is heavily distorted by mirror reversal, wide-angle phone-lens distortion, and lighting. An objective face scan usually shows that people who feel "ugly" have one or two specific below-average metrics — not an overall deficiency — and most of those are improvable. Run the test to see your actual per-feature percentiles instead of guessing.
How ugly am I?
You can't tell "how ugly am I" from a mirror — it shows a reversed version of your face, and phone lenses distort your proportions. The objective answer is a measured one: 17 facial metrics are measured as a per-feature map against population norms, and your separate Face Score percentile is validated against 5,500 human-rated faces (r ≈ 0.8). Almost everyone who asks scores average-or-better on most metrics with just one or two specific weak spots — usually fixable, not a permanent verdict.
Do I look ugly?
Statistically, probably not: most people who take an ugly test out of that worry land mid-range, not at the bottom. The feeling usually comes from mirror reversal, close-range selfie-lens distortion, and comparing your worst lighting to everyone else's best photos. What an honest ugly test can actually do is measure the photo you give it — 17 facial measurements plus a validated Face Score percentile — so "do I look ugly?" or "is my face ugly?" becomes a per-feature read of that picture instead of a feeling, and a low read on a bad photo is flagged as a photo problem, not a face problem.
How do you know if you're actually ugly?
You can't know from a mirror, a selfie camera, or how people treat you — all three are distorted inputs (mirrors reverse you, phone lenses warp proportions at close range, and social feedback is filtered through politeness). The only way to actually tell is to replace the feeling with a measurement: scan a well-lit front-facing photo, look at where each of your 17 metrics lands against population norms, and check your validated Face Score percentile. Almost everyone who asks this discovers they are average-or-better on most measurements with one or two specific weak spots — which is a very different (and more fixable) answer than "ugly."
Am I pretty or ugly?
The binary is the wrong frame — attractiveness is a continuous distribution, and the large majority of people cluster in the middle of it, not at either end. A real answer looks like "63rd percentile, with eye symmetry above median and jaw definition below it," not a pretty/ugly verdict. The test gives you that: a validated percentile for the overall read, plus a per-feature map showing which specific measurements pull you up or down. If you land mid-range — which is the statistically likely outcome — neither label was ever accurate.
Does the am I ugly test work for both guys and girls?
Yes. The same 17 facial-geometry metrics are measured from one front-facing photo regardless of gender, but your results are benchmarked against population norms for your own gender and age group — so a male jawline is compared to male norms and a female jawline to female norms. The validated Face Score percentile works the same way for both. There is no separate "male test" or "female test"; there is one measurement engine with gender-specific reference data.
Is there an "am I ugly" app I need to download?
No download needed — the test runs in your web browser on both phone and desktop. That is deliberate: on desktop the face scan runs on-device in the browser, and on mobile your photo is processed in memory and deleted instantly, which is a stronger privacy setup than installing an app that requests camera-roll access. Open the page, take or upload one photo, and you get your result in about 30 seconds.
Will AI actually tell me if I'm ugly?
It will tell you where you measure — it won't call you ugly, because "ugly" isn't a measurable category. What you get is blunt in a more useful way: a validated Face Score percentile (so you know where a photo of your face lands against 5,500 human-rated faces), and 17 per-feature measurements against population norms, including the ones that land below median. No flattery, no insult — numbers with context. If a metric is weak, it says which one and whether it's the kind that can change.
Can the test rate my side profile?
No — and any tool that claims to score a side profile from a front-facing engine is guessing. This test measures a front-facing photo using a 68-point frontal landmark model, so all 17 metrics (including jaw and chin measurements) are computed from frontal geometry. It does not do dedicated profile scoring of side-view features like nasal dorsum shape or chin projection from the side. If you upload a side profile, the landmark model will either fail to lock or return unreliable numbers, so use a front-facing shot.
Why am I so ugly?
The question usually contains a broken premise. If you feel this way, the most common causes are perceptual, not structural: you see your face 100x more than anyone else does (and in mirror-reverse), your selfies are distorted by close-range wide-angle lenses, and you compare your unedited face to other people's curated photos. That said, the feeling sometimes points at something real — one or two features that genuinely measure below median. The scan separates the two cases: it shows whether your metrics are actually low or whether the problem is the photo and the comparison set. In either case you get specifics instead of a global verdict, and most below-median metrics (skin, definition, symmetry-as-photographed) respond to changes you control.
Can one feature, like my nose, make me ugly?
Rarely. Face perception research consistently finds people read faces holistically — overall harmony and proportion carry more weight than any single feature, which is why plenty of people with prominent noses or asymmetries are widely considered attractive. One feature can pull an overall read down slightly, but it almost never sinks it. The scan is useful here because it quantifies the thing you're fixating on: it shows whether your nose (or jaw, or eye spacing) actually measures outside the typical range, and how the rest of your metrics compare. Fixation usually exaggerates; measurement right-sizes.
Is this an "am I ugly" quiz or an actual test?
It is a measurement, not a quiz. Most "am I ugly quiz" results come from multiple-choice questions you answer about yourself, which just reflects back the self-image you walked in with — if you feel bad about your face, you answer accordingly and the quiz agrees with you. This reads the photo instead: 68 facial landmarks, 17 proportional measurements against population norms, and a Face Score percentile validated against 5,500 human-rated faces. Nothing you believe about yourself changes the number. That is the point — whether you are asking "am I beautiful or ugly" or just want a sanity check, the answer comes from the image rather than from your mood on the day.
Is there an accurate "am I ugly" test?
AI face analysis measures objective metrics — symmetry, proportion, feature placement — more accurately than self-assessment or social feedback. It's not a perfect beauty verdict but it identifies specific low-scoring features and gives actionable data.
Why do I look ugly in photos?
Mainly: mirror reversal (you're used to seeing your flipped face), wide-angle lens distortion at close range, bad lighting creating unflattering shadows, and unnatural posed expressions. All of these are solvable.
What features make a face unattractive?
Measurable factors include significant facial asymmetry, negative canthal tilt, very low facial definition from body fat, poor skin quality, and unflattering lighting in photos. Most of these are improvable.
Can I become more attractive?
Yes. Body fat reduction, improved skin quality, posture correction, and better photo technique all have measurable effects on attractiveness scores. These factors collectively often matter more than fixed bone structure.
What does a face score below 50 mean?
It means certain measured metrics (symmetry, proportion, etc.) score below population average. It doesn't mean you're "ugly" in the full human sense — AI can't measure personality, style, charisma, or non-visual attraction factors.
How does the am I ugly picture test work?
You give it one clear, front-facing picture. A 68-point facial-landmark model (the iBUG 300-W standard) maps the geometry of your eyes, nose, jaw, and facial thirds, then 17 metrics are measured against published population norms by age and gender. You get a reading per feature — a measurement map, not a single "ugly or not" label — plus a separate validated Face Score percentile (checked against 5,500 human-rated faces).
Is the am I ugly test free?
Yes — the picture test and your validated face score are completely free, with no signup (free scans are capped at a real 5 per day; 15 per day with the report, unlimited on Pro). The free scan shows your Face Score percentile and your strongest metric (1 of 17 revealed); the optional full 17-metric report (every measurement, your weakest metric with fixes, and a personalized fix plan) is a one-time fee if you want the detail. Seeing your score is free, and the report is a one-time purchase — no subscription required.
Can I take the am I ugly test with a photo I already have?
Yes. You can upload an existing photo or take a new one — either works. For the most accurate read, use a recent, well-lit, front-facing shot taken at arm's length (60-80cm) so wide-angle lens distortion doesn't skew your proportions. If a feature scores low, the report flags whether it's a photo-quality issue or an actual measurement, so a bad picture won't be mistaken for a real flaw.
Is this "am I ugly" test validated?
Yes, and to our knowledge it is the only "am I ugly" test in the category that publishes a validation. Your Face Score is cross-validated at r ≈ 0.8 against 5,500 human-rated faces in the SCUT-FBP5500 academic benchmark (each face rated by 60 people), so the number tracks how real people rate faces rather than a golden-ratio formula. We also tested our older 17-metric geometry composite, found it did not predict human ratings (r ≈ −0.33), published that result, and stopped using it as a score.
What's the difference between my Face Score and the 17 metrics?
Your Face Score is a validated attractiveness percentile — the single number built to answer "where does a photo of my face land." The 17 metrics (symmetry, canthal tilt, jawline, facial thirds, and so on) are a geometry map: measurements of your individual features against population norms, not a ranking. Read the Face Score for the overall answer, and the 17 metrics to find which one or two features are below median and worth attention.
Am I ugly, or is it just my photo?
Very often it is the photo. Lighting direction alone can change how symmetric a face looks by up to about 23% independent of the actual bone structure (Zaidel & Cohen 2005), and first-impression research finds lighting and pose move the read more than the underlying structure does (Todorov & Porter 2014). Mirror reversal and wide-angle selfie distortion add to it. Take a well-lit, front-facing photo at arm's length (60-80cm) and re-scan before concluding anything about your actual face.
Is this an am-I-ugly quiz or a photo test?
It is a photo test, not a question-based quiz — and that is deliberate. An "am I ugly quiz" asks you to self-report (how you rate your own jawline, how often you get compliments), which just feeds your self-perception back to you. This test measures a photo instead: 17 facial-geometry metrics plus a validated Face Score percentile, so the answer comes from measurement rather than from how you were feeling when you answered. If you searched for a quiz, this is the same question answered with better inputs — one photo, about 30 seconds, no questionnaire.
Is my photo private, or is it stored?
On desktop, analysis runs entirely in your browser. On mobile, your photo is processed in memory by our scan server and deleted immediately after the landmarks are measured — it is never written to disk, never stored, and never used to train anything.
Prefer a human answer? Real people rate your photo — free. Get your Human Score
Further reading
More than one scan
Your results live in a dashboard, free
One tap creates your free account right after your scan — nothing extra to fill in.
8-metric suite
Your complete looksmax scorecard
One AI scan measures all 8 metrics from a single selfie. You've explored one — get the full picture.