Golden ratio face test

The golden ratio is the number φ, approximately 1.618. It appears when a whole quantity relates to a larger part in the same way that the larger part relates to a smaller part. A Golden ratio face analysis applies that comparison to selected distances on a face. It does not claim that every feature should literally equal one universal template. Instead, it asks a narrower question: how close are a few visible proportions in this particular image to 1.618?

Popular diagrams often place a mask, spiral, or grid over a portrait, but those graphics can hide how a result was produced. This Golden ratio face tool uses explicit landmarks and draws each measured segment. You can see the forehead, nose bridge, chin, outer eye corners, mouth corners, nose base, and upper-lip points that support the report. That visibility makes the result easier to inspect and repeat.

Each actual ratio is compared directly with φ. An exact match receives 100 percent. The score decreases in proportion to the absolute distance from 1.618 and is clamped between zero and one hundred. For example, a ratio of 1.338 receives a stronger match than 1.161 because it sits closer to φ. The overall Golden ratio face harmony score is the simple average of the three measurement scores; no secret attractiveness weight is added.

This formula is intentionally readable. A Golden ratio face score does not change because an account purchased credits or used welcome credits. Every valid request uses the same landmarks, constants, quality limits, and rounding rules. Accounts with a successful payment use the configured high-tier OpenRouter model for the written interpretation; other accounts use the low-tier model with a low-tier fallback. That model choice never changes the deterministic ratios or scores.

People are not good at estimating small distance differences from an unmarked portrait. Camera perspective, hair, shadows, expression, and facial asymmetry can make a segment appear longer or shorter. Golden ratio face landmarks replace that guess with repeatable coordinates. The detector identifies hundreds of points, while the report uses a small, documented subset for the three displayed calculations.

The lines also reveal when the input is questionable. If a line lands near hair instead of the forehead, if the mouth is open, or if the head is rotated, the Golden ratio face report should be treated cautiously. Choosing another photo is more honest than interpreting a precise percentage from unstable geometry.

A face is three-dimensional, but a photograph is a two-dimensional projection. Moving a phone closer exaggerates the center of the face. A wide-angle lens can change apparent nose length and outer-face width. Looking up shortens one region while extending another. Smiling changes mouth width. Because a Golden ratio face test measures what the camera records, these changes can alter the ratios without any physical change to the person.

For a useful comparison, place the camera at the same distance, use the same lens, keep eye-level framing, and reproduce the same neutral expression. The Golden ratio face report then becomes a more consistent description of photographic geometry. It still should not be treated as a permanent identity score.

Attractiveness is influenced by far more than three ratios. Symmetry, skin appearance, facial movement, age, cultural context, grooming, hairstyle, expression, warmth, and individuality all affect perception. A Golden ratio face percentage does not measure those qualities. A person can have a lower mathematical match and still be widely perceived as striking, expressive, friendly, or beautiful.

The report therefore avoids labels such as perfect, ideal, ugly, or defective. Golden ratio face language is most responsible when it describes alignment with a number rather than ranking human value. Use the result as a geometry experiment, a photography aid, or a curiosity—not as medical advice or a reason to pursue irreversible changes.

MediaPipe finds the landmarks in the browser. The photograph, normalized coordinates, calculated geometry, and basic quality values are then sent securely to OpenRouter so the AI can write the interpretation. The application does not save the original photo. It stores the compact report needed to make a retry idempotent, and PNG downloading remains local.

The ten-credit price applies only when a valid Golden ratio face report can be returned. No face, poor pose, insufficient image quality, invalid input, AI failure, or report-service failure does not issue a report or charge credits. Purchased and welcome credits use the same geometry; payment history only selects the AI interpretation model tier.