Rather than relying solely on conventional rehabilitation methods, Mustufa decided to use his expertise in artificial intelligence and machine learning to create a solution tailored to his own recovery journey. The result was an AI-powered rehabilitation application called Mirror, a mobile-first web app designed to assist people recovering from Bell’s palsy through guided facial therapy and real-time progress tracking.
The browser-based application transforms a smartphone or laptop camera into an interactive recovery mirror. At the core of the system is Google’s MediaPipe Face Landmarker technology, which is capable of detecting and tracking hundreds of facial landmarks directly within the browser in real time.
Before beginning therapy exercises, the app captures a neutral facial baseline. From there, it monitors facial movements during rehabilitation routines such as smiling, eyebrow raises, and eye-closure exercises. By comparing movements on the left and right sides of the face, the system generates live facial symmetry scores that help users understand their recovery progress more clearly.
Mustufa revealed that OpenAI Codex
played a major role in developing the app’s underlying functionality. Codex was used to help build facial landmark processing, movement analysis, exercise logic, and facial symmetry calculations. Meanwhile, Claude AI
assisted with generating user interface components, layouts, and the app’s mobile-friendly design.
The application also provides visual overlays, percentage-based symmetry indicators, and real-time feedback to make rehabilitation more structured and measurable. Mustufa described the experience as “vibe-coding” the project during recovery, turning a deeply personal health challenge into an innovative AI-driven solution.
The story quickly gained traction among developers and AI enthusiasts online. Many praised Mustufa for transforming a difficult diagnosis into a practical and inspiring technology project. Others shared their own experiences with Bell’s palsy, physiotherapy routines, and recovery timelines, creating a supportive discussion around the condition and the role technology can play in rehabilitation.
Mustufa’s project highlights how modern AI tools are increasingly being used not just for productivity or automation, but also for personal healthcare and recovery support.
