A RAG-powered Snapchat clone built specifically for fitness influencers. SnapConnect leverages cutting-edge AI to revolutionize content creation and personalization in the fitness social media space.
- πΈ Photo/Video capture with disappearing messages
- π± Stories and ephemeral content sharing
- π₯ User authentication and friend management
- π¬ Real-time messaging
- Smart Workout Captions: AI generates personalized captions based on workout type, user history, and personal brand
- Motivational Content: Context-aware motivational posts that align with user's fitness journey
- Workout Plan Generation: Personalized workout routines based on experience level and goals
- Fitness Challenges: Community-engaging challenges tailored to user interests
- Trend-Aware Suggestions: Content recommendations based on current fitness trends
- Content creators building personal brands in fitness
- AI-assisted content that matches personal tone and focus areas
- Engagement optimization through data-driven suggestions
- Seamless integration of fitness knowledge base
- React Native with Expo for cross-platform development
- TypeScript for type safety
- NativeWind (TailwindCSS) for styling
- React Navigation for navigation
- Zustand for state management
- Firebase for authentication, database, and storage
- OpenAI GPT-4 for RAG-powered content generation
- Firestore for user data and content storage
- Firebase Storage for media files
- Node.js 18+ (current version has compatibility issues with Node 16)
- iOS Simulator (for iOS development)
- Android Emulator (for Android development)
- Firebase account
- OpenAI API key
cd SnapConnect
npm install
- Create a new Firebase project at https://console.firebase.google.com
- Enable Authentication (Email/Password)
- Create Firestore Database
- Enable Storage
- Copy your Firebase configuration
- Update
src/services/firebase.ts
with your Firebase config
- Get your OpenAI API key from https://platform.openai.com
- Update
src/services/openai.ts
with your API key
# Start the development server
npm start
# Run on iOS
npm run ios
# Run on Android
npm run android
# Run on web
npm run web
SnapConnect/
βββ src/
β βββ components/ # Reusable UI components
β βββ screens/ # Screen components
β β βββ auth/ # Authentication screens
β β βββ main/ # Main app screens
β βββ navigation/ # Navigation configuration
β βββ services/ # External service integrations
β β βββ firebase.ts # Firebase configuration
β β βββ openai.ts # OpenAI RAG service
β βββ store/ # State management
β βββ types/ # TypeScript type definitions
β βββ utils/ # Utility functions
βββ assets/ # Static assets
βββ docs/ # Documentation
Our RAG (Retrieval-Augmented Generation) system works by:
- Context Building: Gathering user's fitness profile, workout history, and social interactions
- Knowledge Retrieval: Accessing fitness-specific knowledge base and trending topics
- Personalized Generation: Using OpenAI GPT-4 to create content that matches user's:
- Personal brand tone (motivational, educational, casual, professional)
- Focus areas (strength, nutrition, mindset)
- Recent activity and workout history
- Audience engagement patterns
// Generates captions like:
"Just crushed another leg day! πͺ Remember, progress isn't always visible in the mirror, but your body is getting stronger with every rep. Today's focus: building that mind-muscle connection and pushing past comfort zones. What's your favorite way to challenge yourself in the gym? #LegDay #StrengthTraining #FitnessMotivation #GymLife #ProgressNotPerfection"
// Creates personalized motivation based on user journey:
"We all have those days when the gym feels impossible, when motivation is nowhere to be found. But here's the truth: discipline beats motivation every time. Your body doesn't know you're tired - it only knows what you teach it. Every workout completed is proof that you're stronger than your excuses. Keep showing up, even when you don't feel like it. That's where real transformation happens. #Motivation #FitnessJourney #DisciplineOverMotivation"
// Generates structured workout plans:
"Muscle Building - Intermediate Level
1. Barbell Squats: 4 sets x 8-10 reps
2. Romanian Deadlifts: 3 sets x 10-12 reps
3. Bulgarian Split Squats: 3 sets x 12 reps each leg
4. Hip Thrusts: 3 sets x 15 reps
5. Walking Lunges: 3 sets x 20 steps
6. Calf Raises: 4 sets x 15-20 reps
π‘ Focus on controlled movement and mind-muscle connection
π Rest 60-90 seconds between sets"
- β Complete authentication flow
- β Navigation between main screens
- β RAG-powered content generation
- β State management with Zustand
- β Firebase integration ready
- Relevance Score: Generated content relevance to user interests (0-1)
- Engagement Estimation: Predicted engagement based on content analysis
- Brand Alignment: How well content matches user's personal brand tone
- Trend Integration: Incorporation of current fitness trends
src/services/firebase.ts
- Add your Firebase configsrc/services/openai.ts
- Add your OpenAI API keyapp.json
- Update app metadata as needed
- Primary:
#FFDD3A
(Snapchat Yellow) - Dark Background:
#0D0D0F
- Secondary Dark:
#161618
- Accent:
#424242
- Text:
#FFFFFF
/#9E9E9E
- Headers: Bold, high contrast
- Body: Regular weight, good readability
- Captions: Light weight, muted colors
The app is ready for deployment to:
- Expo:
expo publish
- iOS App Store: Build with EAS
- Google Play Store: Build with EAS
- Web:
npm run web
- Model: GPT-4 for high-quality content generation
- Temperature: Varies by content type (0.6-0.8)
- Max Tokens: 150-300 depending on content type
- System prompts optimized for fitness content
- Workout types and exercises
- Fitness motivational phrases
- Trending hashtags and topics
- Nutrition tips and advice
- Camera Integration: Implement actual camera functionality with expo-camera
- Real-time Features: Add Firebase Realtime Database for messaging
- AR Filters: Integrate workout-specific AR effects
- Push Notifications: Implement engagement notifications
- Analytics: Add user behavior tracking
- Content Moderation: Implement AI-powered content filtering
- Target fitness influencers with 1K-100K followers
- Focus on content quality and engagement improvement
- Leverage AI features as key differentiator
- Build community around fitness challenges
- Fork the repository
- Create a feature branch
- Implement your feature
- Add tests if applicable
- Submit a pull request
This project is licensed under the MIT License - see the LICENSE file for details.
Built with β€οΈ for the fitness community
Revolutionizing fitness content creation through AI-powered personalization
Create a .env
file in the root directory with the following variables:
# OpenAI API Key (required for AI features)
EXPO_PUBLIC_OPENAI_API_KEY=your_openai_api_key_here
# Remove.bg API Key (optional - for automatic background removal of AI stickers)
EXPO_PUBLIC_REMOVEBG_API_KEY=your_removebg_api_key_here
To enable automatic background removal for AI-generated stickers:
- Sign up at Remove.bg
- Get your free API key (50 free API calls per month)
- Add it to your
.env
file asEXPO_PUBLIC_REMOVEBG_API_KEY
Benefits:
- Eliminates checkered background patterns on AI stickers
- Creates clean, transparent stickers automatically
- Better integration with photo editing
Note: If you don't set up Remove.bg, the app will still work but AI stickers may have solid backgrounds instead of transparent ones.
SnapConnect now features Advanced AI Sticker Generation with multiple model support and sophisticated prompt engineering:
- DALL-E 3 - Latest OpenAI model with best prompt understanding
- DALL-E 2 - Reliable model with fast generation
- Stable Diffusion XL - Open-source model with artistic styles (coming soon)
- Context-Aware Prompts - Automatically analyzes your photo context
- Fitness-Themed Enhancements - Optimized for fitness content
- Technical Specifications - Professional camera terms and lighting
- Style Optimization - Trending artistic styles and compositions
- Smart Fallback System - Automatically tries alternative models if primary fails
- Background Removal - Automatic transparency using Remove.bg API
- Image Upscaling - Optional quality enhancement (configurable)
- Model Selection UI - Choose your preferred AI model
- Real-time Feedback - See which model generated your sticker
- HD Quality - DALL-E 3 generates 1024x1024 HD images
- Vivid Style - More dynamic and colorful results
- Professional Lighting - Studio-quality lighting specifications
- Clean Edges - Optimized for perfect sticker cutouts
- Contextual Analysis - Detects objects, characters, and abstract elements
- Dynamic Enhancement - Adds appropriate technical specifications
- Fitness Context - Incorporates gym/workout environment awareness
- Artistic Flair - "Trending on ArtStation" style improvements