Eye Tracking Library Tests

Testing various web-based eye tracking solutions for iPad Safari

Session History

🥇 #1 Recommended

GazeCloudAPI

★★★★☆ 4/5 (default)
☆ ☆ ☆ ☆ ☆

High accuracy eye tracking with automatic calibration - More stable than WebGazer

Official API GitHub
✓ Better accuracy
✓ Auto calibration
✓ Low CPU usage
Test GazeCloudAPI
🥈 #2 Needs Work

WebGazer.js

★★☆☆☆ 2/5 (default)
☆ ☆ ☆ ☆ ☆

Scalable webcam eye tracking using MediaPipe Facemesh

Official Site GitHub
✗ Mouse interference
✗ Jumpy tracking
✓ Open source
Test WebGazer.js
Not Yet Implemented

SeeSo / Eyedid SDK

AI-based eye tracking SDK for smartphones, tablets, and webcams. Requires license key.

Official Site GitHub Docs
Not Yet Implemented

MediaPipe + TensorFlow.js

Google's FaceMesh with iris landmark tracking. 478 facial keypoints + iris detection.

TF.js Models NPM Package Blog Post
Not Yet Implemented

Jeeliz FaceFilter + GazeFilter

Lightweight face tracking with accurate pupil position tracking. WebGL-based, low CPU usage.

Official Site GitHub GazeFilter Docs
Not Yet Implemented

face-api.js

Face detection and landmarks on TensorFlow.js. Good for eye position but limited gaze tracking.

Official Docs GitHub
Not Yet Implemented

tracking.js

Lightweight computer vision library with color tracking and face detection (~7KB core).

Official Site GitHub