How does it work?
Pick Analyze with camera or Analyze from image. With the camera, allow access and hold still for about a second; the app auto-captures a frame. With an image, paste (Ctrl+V) or upload a photo. Either way, the app runs Google MediaPipe Face Mesh in your browser: a lightweight model that detects 468 3D facial landmarks (eyes, nose, lips, jaw, etc.) from that frame via JavaScript and WebAssembly. No server uploads. From those points it computes geometric metrics (canthal tilt, fWHR, philtrum-to-chin ratio, gonial angle, and more) in pixel space. Use the Frontal / Side toggle to switch views: capture or upload a frontal shot for frontal metrics, and a separate side shot for profile metrics (gonial angle, facial convexity).