KangooFit is a mobile fitness application designed to transform the way people approach exercise by combining real-time feedback, gamification, and a fully connected ecosystem. The app addresses two of the most common problems in fitness: lack of motivation and incorrect exercise execution.
The application was developed using Android Studio, ensuring a robust and optimized native development workflow. All visual assets and images used throughout the app were generated with Nano Banana from Gemini, providing a consistent and modern visual identity.
Most fitness apps focus on tracking numbers such as steps, calories, or time. KangooFit goes further by actively guiding the user and creating a meaningful reason to stay consistent.
At the center of the experience is a virtual companion. Each user is paired with a kangaroo whose state depends entirely on their physical activity:
- Positive Reinforcement: If the user exercises, the kangaroo becomes happy, gains energy, and grows.
- Accountability: If the user is inactive, the kangaroo loses energy and becomes sad.
This creates an emotional connection that increases accountability and encourages consistency through a gamified experience. In addition, KangooFit includes a competitive leaderboard system with monthly resets. Every month represents a fresh start, giving all users an equal chance to reach the top and stay motivated over time.
- Real-time exercise form detection using the phone camera.
- AI-based validation that ensures users receive points only for correctly executed movements.
- Daily reminders to maintain consistency.
- Global leaderboard with monthly reset.
- Detailed daily and monthly statistics.
- Cross-device experience between phone and smartwatch.
The application is developed natively for Android using Java and XML within Android Studio. Native development was chosen to ensure full control over hardware resources, which is critical for smooth video processing and efficient battery usage.
KangooFit uses Google MediaPipe with the Pose Landmarker model to perform real-time pose estimation.
- The system detects 33 key body points.
- Joint angles are calculated using trigonometry to validate exercise execution.
- All processing is done on-device for low latency and enhanced privacy.
The application integrates with Wear OS to extend functionality to smartwatches.
- Collects passive activity data (steps, calories) via built-in sensors.
- Allows remote exercise selection to trigger the phone camera.
Firebase is used for backend infrastructure:
- Real-Time Database: Enables instant updates across devices and live leaderboard rankings.
- Data Sync: Workout results are uploaded post-session to optimize battery performance.
Google Sign-In is implemented to provide secure, fast, and frictionless user authentication.
- Emotional Gamification: A virtual companion that reacts dynamically to user behavior.
- Fair Competition: AI validation prevents cheating, ensuring points are earned through effort.
- Complete Ecosystem: Seamless connection between mobile and wearable devices.
- Active Coaching: Replaces passive tracking with real-time feedback and guidance.
KangooFit is built with a modular architecture to allow for rapid expansion:
- Support for more complex workouts such as yoga and pilates.
- A customization system for the virtual companion (unlockable items via points).
- Social features including group challenges and team-based competitions.
KangooFit is more than a fitness tracker. It is an intelligent companion that guides users, keeps them accountable, and makes exercise engaging and interactive. The project demonstrates how real-time AI, mobile development, and gamification can be combined to create a meaningful and scalable fitness experience.