
TypeVision
What if a typing trainer could notice where your eyes go?
Role: Frontend Engineer — Design & Implementation
Built a touch-typing trainer that runs in the browser, with JavaScript, optional eye tracking, practice in two languages and progress saved locally. The focused UI lets users train in English and Japanese, track WPM and accuracy in real time, and replay or share challenge runs.
Technical decisions
- MediaPipe FaceMesh runs in the browser, so no video leaves the device
- Eye tracking is optional and calibrated per user; the core trainer works without a camera
- localStorage-first progress keeps the app backend-free and instant
Responsibilities
- Built the game flow for the Practice, Challenge and Progress modes
- Added English/Japanese passage generation and Japanese display options
- Integrated optional MediaPipe FaceMesh eye tracking with calibration and sensitivity controls
- Stored scores, achievements, goals and settings in localStorage
Achievements
- Privacy comes first: core gameplay needs no backend
- A responsive layout, a focus mode, countdowns and live analytics make it easier to use
- Split the app into separate UI, gameplay, storage, audio and translation layers
Technologies
JavaScriptHTMLCSSViteMediaPipe FaceMeshlocalStoragep5.js