Help
Frequently asked questions
The short answers to the questions we are asked most often. If something is missing, write to hello@lookslab.app and we will add it.
The scan
How does the scan actually work?
You upload or capture a front-facing photo and a side photo. LooksLab runs Google’s MediaPipe face mesh model in your browser to find 478 landmark points. Those coordinates are sent to our server (the photo is not), where each derived measurement is scored through a calibrated cubic-bezier curve and aggregated into the four pillar scores. The full breakdown is on the methodology page.
How long does it take?
Most people finish in under a minute. The slowest part is taking the side photo, since people are less practiced at framing their own profile than their front view.
What kind of photo works best?
Even, soft daylight. A neutral, relaxed expression with the mouth closed. Hair pulled away from the face. The camera at eye level and about an arm’s length away. Avoid harsh top lighting, smiles, and selfies taken from below; both distort the underlying geometry.
Why do you need a side photo as well as a front photo?
Half of the most informative facial measurements live in profile rather than front view, including the nasolabial angle, the projection of the chin and forehead, and the mandibular plane. A front photo alone misses them entirely.
Accuracy and limits
How accurate is the analysis?
Landmark detection is reliable within a few pixels on a well-lit face. The measurements derived from those landmarks are accurate to within a couple of percent under good photo conditions. The curve scoring is calibrated against population data, so individual scores have a reasonable margin of uncertainty, particularly at the extremes.
Does lighting matter?
Yes. Harsh shadows under the brow and chin shift several angularity measurements by a meaningful margin. If your score from a top-lit selfie disagrees with your score from a daylight portrait, the daylight portrait is usually the one to trust.
Will my expression change the result?
Yes. Smiling lifts the cheeks, narrows the eyes, and changes the position of the mouth corners. For a stable score, keep your mouth closed and your expression relaxed. Smiling photos are great in life and bad for measurements.
Why do I get a slightly different score each time I scan?
Lighting, angle, focal length, and head tilt all change between photos, and they all feed into the landmark positions. Small variations in any of those produce small variations in the measurements, which then move the score by a point or two. Consistency improves when you scan under the same conditions twice.
Privacy
Where are my photos stored?
Photos are uploaded to a private Supabase Storage bucket as part of the scan flow, for both anonymous and signed-in sessions, so you can revisit your scan and we can render the annotated views. Anonymous scans are scoped to a session-bound Supabase identity. Signed-in scans are scoped to your Google account. Row-level security ensures only your session or account can read your own photos.
Can I delete my photos?
Yes. Open My Scans, find the scan, and delete it. Deletion removes the photos from Supabase Storage and the measurements row from our database. There is no soft delete and no grace period.
Do you train models on my photos?
No. We do not train any model on user photos. The landmark detector is a published Google model that runs in your browser; the scoring server only sees landmark coordinates, not images.
Scores
What is the harmony score?
Harmony measures how well the features of a face relate to each other. It pulls in symmetry, midline alignment, the balance of the facial thirds, and the major intra-feature ratios. It is the closest thing in the report to a single number summary.
What is the angularity score?
Angularity captures bone-driven structure: the gonial angle, the projection of the chin and zygoma, the sharpness of the jaw in profile. Two faces can have similar harmony scores while one reads soft and the other reads chiseled. Angularity is the axis that separates them.
What is the dimorphism score?
Dimorphism measures how strongly features sit on the typical masculine or feminine end of population distributions for your selected gender. Per-gender calibration means a measurement is scored against the modal range for that gender in our calibration set rather than averaged across the full population.
What are the Health Indicators?
A small set of measurements that have been associated with general health and developmental signals in the literature. They are correlates, not diagnostics. A high or low score here is not a clinical claim about anyone’s health.
Is the golden ratio actually predictive of attractiveness?
Partly, with a lot of nuance. Geometric harmony measurements, including some that are close to the golden ratio, do correlate with rated attractiveness in published studies, but the effect is modest and the variance unexplained by geometry is large. Attractiveness is also shaped by familiarity, culture, fashion, expression, and individual taste. Treat any geometric score as one input among many.
Account and payment
Do I need to sign in to use LooksLab?
No. You can run a scan without signing in and see a preview of your top strengths and weaknesses. Signing in lets you save scans across devices, return to them later, and manage your subscription.
How does pricing work?
You can try the scan without paying. Anyone gets a free preview of their top strengths and weaknesses. The four pillar scores, the overall score, and the full per-metric breakdown unlock with a subscription.
What does paying unlock?
Every measurement, every per-curve score, the full annotated front and side photos, and the saved per-metric history of every scan on your account. The subscription comes in three intervals (monthly, quarterly with a 20 percent saving, and yearly with a 60 percent saving) and can be canceled at any time from the Stripe customer portal.
Can I cancel and get a refund?
You can cancel your subscription at any time from the Stripe customer portal, which stops future renewals while preserving access through the end of the current period. Refund requests are reviewed case by case within 14 days of the most recent charge. If the report failed to render after a successful charge, write to legal@lookslab.app and we will fix it and refund the affected payment. Refund requests based on disagreement with your score are not granted.
Technical
Can I use LooksLab on a phone?
Yes. The full scan flow runs on modern iOS and Android browsers. We test against the most recent two major versions of Safari and Chrome on each platform.
Which desktop browsers work?
Recent versions of Chrome, Edge, Firefox, and Safari. The landmark detector uses WebAssembly and WebGL, both of which need to be enabled, which is the default in every modern browser.
Does it work offline?
No. The landmark model is downloaded the first time you scan and then cached, but scoring runs on our server, and photo storage and saved scans both go over the network. A scan needs an internet connection end to end.