the global rise in lifestyle-related diseases has intensified the demand for intelligent, accessible, and personalized digital health interventions. While numerous mobile fitness applications exist, the majority suffer from critical limitations: rigid meal logging workflows, absence of real-time AI coaching, poor offline functionality, and a lack of integration between nutrition tracking, physical planning, and behavioral feedback. This paper presents BFIT, a novel AI-driven Android application that addresses these deficiencies through a unified, modular framework combining on-device machine learning, large language model (LLM)-based conversational coaching, and a hybrid local-cloud data architecture. The system integrates Google ML Kit for real-time barcode-based nutrition scanning and image-based meal recognition, the Gemini Generative AI SDK for context-aware dietary and fitness coaching, and a dual-layer persistence model using Room (local) and Firebase Firestore (cloud). A BMI-driven personalized planner generates user-specific bulk, lean, or maintenance programs. Preliminary functional evaluation demonstrates that BFIT achieves accurate product nutrition retrieval via OpenFoodFacts API integration, responsive AI coaching interactions, and reliable weekly progress analytics through structured weight logging and MPAndroidChart visualizations. The proposed architecture establishes a reproducible blueprint for the next generation of intelligent mHealth applications.
Pushp Raj (Sun,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: