left drawer
profile screen
home screen
cook mode
items list

🍽️ NextMeal

Your intelligent cooking companion that transforms recipe chaos into culinary confidence

💡 Inspiration

The kitchen should be a place of creativity and joy, not frustration and confusion. We've all been there: juggling multiple recipe tabs, losing track of ingredients, and struggling to follow cooking instructions while our hands are covered in flour.

The inspiration for NextMeal came from a simple observation: cooking is hard not because of the recipes themselves, but because of how we manage them. We envisioned an app that would:

Eliminate the friction of recipe management
Bring AI intelligence to your cooking experience
Work seamlessly whether you're online or offline
Respect your privacy while offering cloud sync when you need it

We wanted to build something that feels like having a personal sous chef in your pocket—one that remembers your preferences, answers your questions, and makes cooking accessible to everyone.

🎯 What It Does

NextMeal is a comprehensive recipe management platform that combines the power of AI with intuitive design to revolutionize your cooking experience.

Core Capabilities

🎥 Intelligent Recipe Extraction

Paste any video link from YouTube, Instagram, or TikTok
AI-powered extraction automatically pulls ingredients and instructions
Smart parsing handles multiple formats and recipe structures
Manual entry option for traditional recipes or personal creations

👨‍🍳 Interactive Cook Mode

Step-by-step guidance with a distraction-free interface
Ingredient checklist to track what you've added
Progress tracking so you never lose your place
Hands-free design optimized for kitchen use

🛒 Smart Grocery Lists

Automatic consolidation of ingredients across multiple recipes
Intelligent grouping by category
Swipe-to-delete for quick list management
Real-time sync across devices (for signed-in users)

💬 AI Recipe Assistant

Ask questions about substitutions, techniques, or timing
Get personalized suggestions based on your dietary preferences
Modify recipes on the fly with AI guidance
Context-aware responses that understand your cooking situation

🔐 Flexible Authentication

Start as guest with full local functionality
Upgrade to Google Sign-In for cloud sync
Seamless account linking preserves all your data
Privacy-first approach with anonymous mode

🌓 Beautiful Theming

Light and dark modes for any environment
Consistent design language across all screens
Smooth animations and transitions
Accessibility-focused interface

🛠️ How We Built It

Technology Stack

Frontend Framework

Flutter for cross-platform development (iOS, Android, Web)
Dart as the primary programming language
Material Design 3 for modern UI components

State Management & Architecture

Riverpod for reactive state management and dependency injection
GoRouter for declarative navigation with auth-based redirects
Clean Architecture with feature-based modular structure

Backend & Services

Firebase Authentication for secure user management
Firebase Analytics for usage insights and optimization
Google Gemini AI for recipe extraction and chat functionality
SQLite for local data persistence
Firestore for cloud synchronization

Key Packages & Libraries

dependencies:
  flutter_riverpod: ^2.4.0      # State management
  go_router: ^12.0.0            # Navigation
  sqflite: ^2.3.0               # Local database
  firebase_auth: ^4.15.0        # Authentication
  firebase_analytics: ^10.7.0   # Analytics
  google_sign_in: ^6.1.5        # Google OAuth
  google_generative_ai: ^0.2.0 # Gemini AI
  shared_preferences: ^2.2.2    # Local storage

Architecture Highlights

Feature-Based Structure

lib/
├── core/                    # Shared utilities
│   ├── database/           # SQLite helpers
│   ├── router/             # Navigation config
│   ├── sync/               # Cloud sync engine
│   ├── theme/              # Theming system
│   └── widgets/            # Reusable components
├── features/               # Feature modules
│   ├── add_recipe/         # Recipe extraction
│   ├── auth/               # Authentication
│   ├── cook_mode/          # Cooking interface
│   ├── grocery_list/       # Shopping lists
│   ├── home/               # Recipe library
│   └── recipe_detail/      # Recipe view & chat
└── services/               # Business logic
    ├── analytics/          # Event tracking
    ├── auth/               # Auth management
    └── subscription/       # Feature gating

Sync Architecture

We built a sophisticated sync engine that handles:

Conflict resolution with timestamp-based merging
Offline-first design with queue-based sync
Automatic retry with exponential backoff
Migration service for guest-to-authenticated transitions

AI Integration

Our extraction pipeline uses:

Multi-stage parsing with fallback chains
JSON validation and error recovery
Context-aware prompting for better accuracy
Performance monitoring for optimization

🚧 Challenges We Ran Into

1. AI Extraction Reliability

Challenge: Video recipe formats vary wildly—some have structured descriptions, others have chaotic formatting.

Solution: Built a multi-stage extraction pipeline with:

Primary extraction using Gemini AI
Fallback parsers for common formats
JSON validation and repair mechanisms
Graceful degradation to manual entry

2. Offline-First Sync

Challenge: Users need the app to work without internet, but also want cloud sync when available.

Solution: Implemented a queue-based sync system that:

Stores operations locally when offline
Automatically syncs when connection returns
Resolves conflicts using timestamp logic
Provides visual feedback on sync status

3. Guest-to-Authenticated Migration

Challenge: Preserving user data when upgrading from guest to Google account.

Solution: Created a migration service that:

Uploads all local data to Firestore
Links anonymous account to Google account
Maintains data integrity during transition
Provides progress feedback to users

4. Usage Limit Tracking

Challenge: Implementing fair usage limits for free vs. pro tiers with daily resets.

Solution: Built a usage tracking service with:

Lifetime counters for free tier
Daily reset logic for pro tier (24-hour windows)
Atomic operations to prevent race conditions
Clear visual feedback on remaining usage

5. Cross-Platform Video Extraction

Challenge: Different video platforms (YouTube, Instagram, TikTok) have different metadata structures.

Solution: Developed platform-specific extractors with:

URL pattern detection
Platform-specific API calls
Unified data model for all sources
Fallback to generic extraction

🏆 Accomplishments That We're Proud Of

✨ Technical Achievements

Robust Sync Engine
Built a production-ready sync system that handles offline scenarios, conflicts, and migrations seamlessly.
AI-Powered Extraction
Achieved 85%+ accuracy in recipe extraction across diverse video formats.
Comprehensive Test Coverage
Wrote 150+ unit and integration tests covering critical paths and edge cases.
Clean Architecture
Maintained strict separation of concerns with feature-based modules and dependency injection.
Performance Optimization
App launches in under 2 seconds and handles 1000+ recipes without lag.

🎨 User Experience Wins

Seamless Onboarding
Users can start cooking in under 30 seconds without creating an account.
Intuitive Cook Mode
Distraction-free interface with large text and easy navigation for kitchen use.
Smart Grocery Lists
Automatically consolidates ingredients and eliminates duplicates.
Responsive Design
Works beautifully on phones, tablets, and web with adaptive layouts.

🔒 Privacy & Security

Privacy-First Design
Full functionality available without sign-in or data collection.
Secure Authentication
Firebase Auth with industry-standard OAuth and anonymous mode.
Data Ownership
Users can export or delete their data at any time.

📚 What We Learned

Technical Lessons

1. AI Integration is an Art

Working with Gemini AI taught us that prompt engineering is crucial. We learned to:

Provide clear, structured prompts
Handle hallucinations and errors gracefully
Validate AI outputs rigorously
Build fallback mechanisms

2. Offline-First is Complex

Building a true offline-first app requires:

Queue-based architectures for pending operations
Conflict resolution strategies that users understand
Visual feedback on sync status
Extensive testing of edge cases

3. State Management Matters

Riverpod's dependency injection and reactive updates made complex state management elegant:

Providers compose beautifully
Testing becomes straightforward
Code remains maintainable at scale

4. Testing Saves Time

Investing in comprehensive tests early paid dividends:

Caught bugs before production
Enabled confident refactoring
Documented expected behavior
Reduced debugging time

Design Lessons

1. Simplicity Wins

Users don't want complexity—they want to cook. We learned to:

Remove unnecessary steps
Provide sensible defaults
Hide advanced features until needed
Focus on the core workflow

2. Feedback is Essential

Every action needs immediate visual feedback:

Loading states for async operations
Success/error messages
Progress indicators
Sync status indicators

3. Accessibility Matters

Designing for accessibility improves the experience for everyone:

Large touch targets for kitchen use
High contrast for readability
Clear typography
Logical navigation flow

🚀 What's Next for NextMeal

Short-Term Goals (Next 3 Months)

🎤 Voice Control

Hands-free navigation during cooking
Voice commands for timers and steps
Read-aloud instructions for accessibility

📱 Social Features

Share recipes with friends and family
Recipe collections and favorites
Community ratings and reviews

🔔 Smart Notifications

Cooking reminders based on meal planning
Ingredient expiration alerts
Recipe suggestions based on available ingredients

Medium-Term Goals (6-12 Months)

🍎 Nutrition Tracking

Automatic calorie calculation
Macro and micro nutrient breakdown
Dietary restriction filtering
Meal planning integration

🤖 Advanced AI Features

Recipe generation from available ingredients
Personalized recommendations based on taste preferences
Cooking skill assessment and adaptive instructions
Ingredient substitution suggestions

💳 Monetization

RevenueCat integration for subscription management
Pro tier features (already architected)
Family sharing plans
Lifetime purchase option

Long-Term Vision (1-2 Years)

🌍 Global Expansion

Multi-language support (Spanish, French, German, etc.)
Regional recipe databases
Cultural cooking guides
International ingredient databases

🏪 Grocery Integration

Direct ordering from grocery delivery services
Price comparison across stores
Automatic coupon application
Inventory management with smart home integration

📊 Analytics Dashboard

Cooking statistics and achievements
Cost tracking per recipe
Time management insights
Sustainability metrics (food waste reduction)

🎓 Educational Content

Cooking technique videos
Knife skills tutorials
Ingredient guides
Chef masterclasses

🎨 Design Philosophy

Principles

Simplicity First
Every feature must justify its existence. If it doesn't make cooking easier, it doesn't belong.
Offline-First
The app should work perfectly without internet. Cloud sync is a bonus, not a requirement.
Privacy-Respecting
Users own their data. We collect only what's necessary and provide full control.
Accessible to All
Cooking should be accessible regardless of skill level, device, or ability.
Delightful Interactions
Small animations and thoughtful feedback make the experience memorable.

Development Setup

# Clone the repository
git clone  https://github.com/aankit-roy/NextMeal

# Install dependencies
flutter pub get

# Run tests
flutter test

# Run the app
flutter run