REFRIGERATOR ASSISTANT MOBILE APPLICATION FOR FOOD MANAGEMENT, RECIPE SUGGESTIONS FROM INGREDIENTS AND IMAGES, AND MEAL PLANNING.

Abstract

This thesis presents the design and development of a household-oriented refrigerator management application built using Flutter and the Firebase ecosystem. The system aims to support families in organizing and monitoring their food supplies, reducing waste, coordinating shared refrigerators, and improving daily meal planning. The application provides an intuitive, user-friendly interface that enables family members to collaboratively update ingredient quantities, track real-time changes across multiple devices, manage multiple shared fridges, and receive timely notifications regarding important updates. The methodology incorporates key Firebase services–including Firestore, Authentication, and Cloud Messaging–to deliver reliable real-time synchronization and cross-device data consistency. For media storage, the system uses Cloudinary to store and deliver images, while only image URLs are stored in Firestore. In addition, the system integrates Gemini AI to enhance intelligence-driven functionalities such as normalizing ingredient names and correcting minor misspellings, performing semantic ingredient matching, suggesting meals based on available ingredients, and recognizing dishes from user-submitted images. Moreover, the system can generate cooking recipes from selected ingredients and produce illustrative dish images to make the cooking guidance easier to follow. The input suggestion mechanism also helps users type faster and still receive relevant recommendations even with minor spelling errors. Experimental results indicate that the application effectively improves household coordination, enhances the utilization of existing ingredients, and provides users with accurate and practical meal suggestions. In addition, expiry tracking and timely reminders help reduce the risk of consuming expired food and support safer and healthier eating habits. In conclusion, this project demonstrates that integrating Flutter with Firebase and Gemini AI can produce a practical, scalable, and user-centric solution for managing food inventories in domestic environments. The system not only supports daily cooking activities but also lays the groundwork for future expansion toward smart-home ecosystem integration.