WebSight: A Semantic Gaze-to-Code Mapping Layer for Web Experiments
By Ningzhi Tang, SaNDwich Lab
WebSight maps eye-gaze or mouse coordinates to the corresponding line, column, token, and syntax context in real time.
Mapping remains aligned while the code is highlighted, scrolled, or edited—without screenshot-based token detection, OCR, or manual remapping.
Gaze information
- Data source: Starting…
- Screen position: X: 0, Y: 0
- Viewport position: X: 0, Y: 0
- Code position: line 0, column 0
- Token:
- Syntax context: