A beautiful, modern web interface for GrannyGPT - your wise Caribbean grandmother AI that speaks authentic Creole and shares cultural wisdom, stories, and advice.
- Authentic Caribbean Creole AI - Trained on 6,000 Caribbean Creole conversations
- Beautiful Modern Interface - Warm, retro-inspired design with smooth animations
- Light & Dark Mode - Toggle between themes with smooth transitions
- Smart Conversation Management - Create, switch, and delete chat sessions
- Configurable Memory - Adjust conversation history limit (10-50 messages)
- Streaming Responses - Real-time AI responses with typing indicators
- Responsive Design - Works perfectly on desktop and mobile
- Production Ready - Professional, polished interface
- Node.js 22+ (LTS recommended)
- npm or yarn
-
Clone the repository
git clone https://github.com/RolexAlexander/GrannyGPT.git cd GrannyGPT -
Install dependencies
npm install
-
Configure environment
cp .env.example .env
Edit
.envand add your configuration:VITE_API_URL=https://your_hosted_version_of_model_url VITE_MODEL_NAME=hf.co/RolexAlexander/GrannyGPT-3.2-Carib:F16
-
Start development server
npm run dev
-
Open your browser Navigate to
http://localhost:5173
Create a .env file based on .env.example:
| Variable | Description | Example |
|---|---|---|
VITE_API_URL |
Your hosted model API endpoint | https://783a513db26e.ngrok-free.app/api/chat |
VITE_MODEL_NAME |
HuggingFace model identifier | hf.co/RolexAlexander/GrannyGPT-3.2-Carib:F16 |
- Conversation Memory: Adjustable from 10-50 messages in settings
- Theme: Light/Dark mode toggle in hamburger menu
- Auto-scroll: Automatic scrolling to latest messages
- Message Fade: Older messages fade for better focus
- Warm Color Palette: Caribbean-inspired orange and green gradients
- Glassmorphism Effects: Modern translucent UI elements
- Smooth Animations: Floating elements and transitions
- Responsive Layout: Works on all medium and large screen sizes
- Smart Message History: System maintains conversation context
- Typing Indicators: Shows when GrannyGPT is responding
- Message Timestamps: Track conversation timing
- Chat Management: Create, switch, and delete conversations
- Streaming Responses: Real-time message display
GrannyGPT is powered by a fine-tuned language model trained specifically on Caribbean Creole conversations.
- Model: GrannyGPT-3.2-Carib
- Training Dataset: 6K Caribbean Creole Conversations
- Language: Authentic Caribbean Creole
- Personality: Warm, wise Caribbean grandmother
- Specialties: Cultural wisdom, cooking advice, life stories, traditional knowledge
- Frontend: SvelteKit
- Styling: Tailwind CSS
- Build Tool: Vite
- Language: JavaScript
- State Management: Svelte Stores
- API: Streaming JSON response handling
# Development
npm run dev # Start dev server
npm run build # Build for production
npm run preview # Preview production build
# Code Quality
npm run check # Type checking
npm run lint # ESLint
npm run format # Prettier formattingGrannyGPT/
βββ src/
β βββ lib/
β β βββ stores/
β β βββ chat.js # Chat state management & API calls
β βββ routes/
β β βββ api/
β β β βββ chat/
β β β βββ +server.js # Streaming API endpoint handler
β β βββ +page.svelte # Main application interface
β βββ app.css # Global styles & Tailwind imports
β βββ app.html # HTML template
βββ static/ # Static assets (logo.png)
βββ tailwind.config.js # Tailwind CSS configuration
βββ .env.example # Environment variables template
βββ package.json # Project dependencies
npm run buildThe app is ready for deployment on any SvelteKit-compatible platform:
- Vercel: Connect your GitHub repo for automatic deployments
- Netlify: Connect repo or deploy build folder
- Railway: Deploy with Node.js preset
- Self-hosted: Use SvelteKit Node adapter
- Fork the repository
- Create your feature branch (
git checkout -b feature/AmazingFeature) - Commit your changes (
git commit -m 'Add some AmazingFeature') - Push to the branch (
git push origin feature/AmazingFeature) - Open a Pull Request
This project is licensed under the MIT License - see the LICENSE file for details.
- π Repository: GitHub
- π€ Model: HuggingFace
- π Dataset: Kaggle
If you have questions or need help:
- Check existing Issues
- Create a new issue with detailed information
- Review the configuration section above
Made with β€οΈ for the Caribbean community
"Code wid purpose, not just code fi code." - GrannyGPT
