The scale
The scale sets the range and groups scores into bands. Use the table below to see how this site displays a score. Other sites can use different cutoffs and names.
Explore your facial proportions with a free, private photo scan. Get an optional AI report for a 0–8 score, feature explanations and styling guidance.
See Mini and Complete sample reports
478 SAMPLE LANDMARKS1.16 : 1
The gap between the inner corners divided by the average visible eye width.
WHAT IT MEANS
It is an informal framework for describing facial appearance on a numeric range. This tool interprets one submitted photo on a 0–8 scale.
The scale sets the range and groups scores into bands. Use the table below to see how this site displays a score. Other sites can use different cutoffs and names.
The AI looks at the visible structures and overall balance in your image. Its comments depend on what the lens, crop and light let it see.
Your report shows the score and its band alongside the clarity of the photo, the model’s observations and concrete tips for a retake.
PHOTO GUIDE
The PSL Scale reads what is visible in one picture. A repeatable setup gives you more useful observations and makes retakes easier to compare.
Rest your phone on a stable surface or ask someone to hold it. Leave some distance between the lens and your face. Cropping a wider frame is usually a better starting point than holding a wide-angle camera very close. Keep the top of your head and your chin inside the picture.
Choose or drop a JPG, PNG or WebP up to 5 MB. Check the dimensions and the exposure hint. If a strong side light hides details, face a window or use a diffuse lamp. Use a clear original with your forehead and chin in frame.
Open AI analysis, confirm the photo is yours and that you are at least 18, then start the analysis. Read the score band alongside the observations and photo confidence. Try one retake suggestion at a time to see what changed.
REPORT COMPARISON
Start with the measurements. Use credits when you want an interpretation of your own photo. Failed generations leave your balance unchanged.
Generated by the report model using fictional adult portraits. Switch tiers to explore the findings and actions included.
Move the light source slightly to the side to add subtle shadow and restore cheek contour definition.
How to compare Keep the same camera position and head angle; compare the original with a retake using side lighting to see how the cheek contours become more distinct.
2026-11-18 · mistralai/mistral-small-2603
| What you get | Free scan | Mini · 2 cr | Complete · 5 cr |
|---|---|---|---|
| Face mesh | On device | On-device scan remains available | On-device scan remains available |
| Measurements | Four image-space measures | Local measures plus context where visible | Local measures plus limitations |
| Score and band | — | 0–8 photo estimate | 0–8 photo estimate |
| Feature observations | Exposure and framing guidance | Two facial findings, one photo condition and a retake comparison | Eight named assessments with confidence |
| Style and action plan | — | Retake tips | 3–5 options and a 30-day plan |
| Export and review | Download measurements without the photo | Private saved report | Private report, print or save as PDF |
| Action | Credits | When it applies |
|---|---|---|
| Free scan | 0 | Any eligible local photo, no account needed |
| Unlock Mini Insight | 2 | First successful Mini for this scan |
| Unlock Complete directly | 5 | First successful Complete for this scan |
| Upgrade your Mini | 3 | Complete on the same saved scan |
| Reopen a saved report | 0 | From My reports in your account |
SCORE GUIDE
Use the band to locate the number on this site’s range. Read it alongside the report and the quality of the picture.
These bands organize this site’s estimates. They carry no population percentiles.
| Score | Band | How to read it |
|---|---|---|
| 0.0–2.0 | Low band | Read the photo confidence and observations before interpreting the number. |
| 2.1–3.9 | Lower midrange | Use the visible-feature notes to see what the model weighted. |
| 4.0–5.4 | Midrange | Compare the observations and photo conditions before comparing scores. |
| 5.5–6.9 | Upper midrange | Look for consistent observations across photos taken under similar conditions. |
| 7.0–8.0 | High band | Read the number as this site’s upper range for the image, with the same photo and model limitations. |
| View | Output |
|---|---|
| On-device check | File compatibility, dimensions and approximate exposure. |
| Sample mesh | Detected positions and visible ratios for an example photo. |
| AI analysis | Photo estimate and band, clarity, observations and retake tips. |
| What you see | Try this next |
|---|---|
| Uneven shadow | Face a diffuse light to make both sides visible. |
| Close wide-angle crop | Move the camera back, then crop the frame if needed. |
| Turned or tilted head | Set the camera at eye level and face it directly. |
READING YOUR RESULT
Read the report alongside the photo you sent. If it calls out uneven lighting, check the shadows before you interpret comments on visible balance. If the model says the outline was obscured, clear your hair away from the jaw or move the camera slightly farther back.
Try changing one condition at a time. Save the camera position, retake the photo with softer light, and compare what the two images show. Small score differences can reflect the camera and the model’s judgment. Look for observations you can see for yourself in both photos.
SAMPLE REPORTS
Explore how light and framing shape the visible observations. Each example uses a fictional adult portrait and an illustrative result.
Illustrative portraitUse a window or diffuse light in front of you. A neutral expression and clear edges make this a useful baseline. Keep this setup if you plan to compare a second photo.
View sample report
Illustrative portraitMove the camera farther back and crop afterward if needed. This helps when an arm’s-length picture makes central features look noticeably different from your usual photos.
View sample report
Illustrative portraitFace the light source or add soft light from the darker side. Keep the crop and expression similar so you can see what the change in lighting did to the picture.
View sample reportThe PSL Scale is an informal online convention for discussing facial appearance. This tool uses a 0–8 range for an AI estimate of one submitted image. A PSL Scale report also points out the visible details and photo conditions that informed the estimate. Different communities and products can assign different meanings to the same number.
The AI report rounds the image estimate to one decimal and looks up its band in the rating chart on this page. The band depends only on that displayed estimate. We do not use a database of people to produce a rank or a percentile.
You can open a photo and inspect its exposure, detected landmarks, eye spacing, face shape ratios and mirror deviation. You can also switch between photo measurements and a depth projection on the example portraits. Those measurements apply to the example portraits. An AI report requires the separate submission step.
Camera distance, exposure, head tilt, crop and expression can all change what the model sees. Start by checking the observations and confidence label for each photo. If you want a comparison, use the same device from the same distance, with eye-level framing and similar lighting. Change one condition at a time.
A PSL Scale result is for informational and entertainment use. It comes from a vision model interpreting a two-dimensional image without a validated attractiveness scale or reference population. Confidence describes the clarity of the submitted photo. The tool has no way to measure someone’s social appeal or predict how others will respond to them.
Take a front-facing photo with a relaxed expression, the whole face in frame and enough light to see both sides. Keep hands, hair and anything else that covers key features away from the face. Start with an unfiltered image from eye level. A full-resolution original is usually preferable to a screenshot of a heavily compressed photo.
Your browser prepares a smaller copy first. When you choose Analyze my photo, that copy is sent through our server to OpenRouter and its model provider for processing. The server keeps the report text in your account. Your original and re-encoded photo stay out of site storage. The privacy policy explains the temporary browser draft and how to delete saved reports.
The on-device scan and examples work without an account. Sign in with a verified email before using credits for an AI report. The comparison table gives the exact cost of Mini, Complete and the same-photo upgrade. If your balance is too low, review credit pack availability on the pricing page.
Use your own photo and confirm you are at least 18. A group photo, celebrity headshot, heavily filtered picture or obscured face gives the model limited information about the photo task. If you are curious about the output without using your own image, start with the labeled fictional examples.
Start with the free estimate and a few practical pointers for a clearer retake.