Liora - Private Mobile Application

Project Overview

Liora is a sophisticated, enterprise-grade private mobile application designed to provide a holistic wellness experience. It combines advanced menstrual health tracking with a seamless integrated marketplace. Built using the Flutter framework and Dart, Liora leverages high-performance state management, a robust cloud backend, and on-device Machine Learning (TensorFlow Lite) to ensure a secure, private, and highly personalized user experience.

Application Purpose

Liora is engineered to empower users by providing:

• Intelligent Tracking (AI-Powered): Data-driven insights into menstrual cycles using personalized ML models (Accuracies >75%).
• Wellness Marketplace: A curated shop for health and wellness products with integrated order management.
• Data Privacy (Privacy-First): All sensitive health predictions and ML inference happen 100% locally on-device.
• Administrative Control: A comprehensive backend suite for managing users and inventory.
• Personalized Recommendations: Dynamic diet and wellness advice based on current cycle phases.

Platform Support

Technology Stack

Liora utilizes a modern, scalable technology stack:

• Core Framework: Flutter (SDK ^3.19.x)
• Programming Language: Dart
• Machine Learning: TensorFlow Lite for on-device inference.
• Backend-as-a-Service (BaaS): Firebase
  • Authentication: Secure user identity management.
  • Cloud Firestore: Real-time NoSQL database.
  • Cloud Storage: Media assets and user profile storage.
• State Management: Provider
• Key Dependencies:
  • tflite_flutter: For on-device ML model execution.
  • table_calendar: For intuitive cycle visualization.
  • cached_network_image: For optimized shop asset loading.
  • google_fonts: For premium typography.
  • shared_preferences: For local persistence of user settings.

Architecture Overview

The project follows a Modular Layered Architecture, emphasizing separation of concerns and maintainability. By decoupled business logic from the UI, the application ensures high testability and scalability.

• Presentation Layer: State-aware widgets (Provider-based) and screen compositions.
• Business Logic Layer (Services/AI): Providers, Service classes, and ML Inference Service handling data calculations.
• Data Layer (Models): Strongly typed Dart models for consistent data handling (e.g., MLCycleData).
• Core Layer: Shared utilities, themes, and application-wide configurations.

Project Folder Structure

LIB Directory Breakdown

Below is the structural representation of the core application logic:

lib/
├── admin/          # Administrative dashboards and management tools
├── core/           # App-wide constants, themes, and session management
├── home/           # Main dashboard and enhanced cycle algorithms
├── models/         # Data structures (Cart, Products, MLCycleData, etc.)
├── onboarding/     # User entry flow and welcome screens
├── screens/        # Authentication, AI settings, and insights panels
├── services/       # ML Inference, AI services, and business logic
└── shop/           # Marketplace interface and commerce features

Module Descriptions

Installation & Local Development Setup

Prerequisites

• Flutter SDK: ^3.19.x
• Dart SDK: Compatible with the installed Flutter version.
• Python (Optional): Required only if re-training ML models using train_cycle_model.py.
• Android Studio / VS Code: With Flutter and Dart plugins.

Step-by-Step Setup

1. Clone the Repository:

   git clone <repository-url>
   cd liora

2. Install Dependencies:

   flutter pub get

3. Download/Train ML Model:

   • Run python train_cycle_model.py to generate the .tflite model.
   • Place the model file in the assets/ directory.

4. Configure Firebase:
   As this is a private project, you must provide your own google-services.json (for Android) and update firebase_options.dart.

AI Implementation Guides

For detailed technical information on the AI/ML system, refer to the following local documents:

• Quickstart Checklist - Start Here
• Technical Roadmap - Feature implementation status.
• AI Integration Guide - Service-level details.
• ML Architecture Reference - Neural network specifics.
• Implementation Summary - Project overview.

Running the Application

Using Terminal

flutter run

Android Build Instructions

To generate a signed APK:

flutter build apk --release

Usage Guidelines

Navigating the Codebase

• Business Logic: Located in lib/services/. Do not place complex logic inside Widget files.
• UI Components: Reusable widgets should be modularized. Primary screens reside in lib/screens/ and feature-specific folders (e.g., lib/shop/).
• State Management: Use Provider.of<T>(context) or Consumer<T> to access application state.

Extending Features

• Adding a New Screen: Create the file in lib/screens/ or a new module folder. Ensure it follows the established app_theme.dart for visual consistency.
• Adding a Module: Maintain the existing folder structure. If a new domain is introduced (e.g., Settings), create a corresponding folder in lib/.
• Consistency: Always use the defined Models for data piping to ensure type safety across the application.

Contribution & Leadership

Team Lead

• Alwin Madhu - Technical Lead & Architect

Development Team

Security & Access Policy

• Developer Access: Only authorized developers listed in the Contribution section are granted access to this repository.
• Data Privacy: Health data is encrypted at rest and processed locally for ML predictions.
• Local Development: Authorized developers must use their own development environment keys and comply with internal security guidelines.

Future Roadmap

• AI Insights: Enhanced cycle prediction using machine learning patterns.
• Diet Recommendations: Personalized phase-specific nutrition advice.
• Deployment: Google Play Store and Apple App Store release.
• iOS Support: Fine-tuning UI/UX for iOS standards.
• Notifications: Push notifications for cycle reminders.

License & Usage Restrictions

Usage of this software is governed by a private license. All rights reserved. No part of this application may be reproduced or transmitted in any form without the express written permission of the Team Lead, Alwin Madhu.

Version Information

Current Version: 1.2.0
Last Updated: March 2026
Status: Alpha Release (AI/ML Integrated)