Pickly is an intelligent decision-making assistant that helps you make better choices in everyday situations. Whether you're deciding what to cook for dinner, what activity to do, or need help with any decision, Pickly uses AI to analyze your context and provide personalized suggestions.

Decision fatigue is real. Pickly was built to help people make faster, better decisions in everyday situations—from figuring out what to cook with available ingredients to choosing spontaneous activities.

• Integrating multiple AI models for different tasks (visual analysis, conversation, recommendations)
• Implementing real-time conversational AI with context awareness
• Designing secure database schemas with Row Level Security
• Building responsive UIs that handle image uploads and AI responses seamlessly

Built with React 18, TypeScript, and Tailwind CSS on the frontend. Backend powered by Lovable Cloud (Supabase) with PostgreSQL, authentication, and Edge Functions. Multiple AI models (Gemini and GPT) handle visual analysis, conversations, and recommendations.

• Balancing AI model selection for cost, speed, and accuracy
• Achieving reliable ingredient detection from varied fridge photos
• Creating context-aware conversations without repetition
• Ensuring data security while maintaining fast performance

Ask any decision-related question and get AI-powered guidance through an interactive conversation that understands context and asks follow-up questions when needed.

Upload a photo of your fridge contents and receive personalized recipe suggestions based on available ingredients, dietary preferences, and cooking time.

Get spontaneous activity suggestions tailored to:

• Current time of day
• Weather conditions
• Your personal preferences
• Available resources

Secure user accounts with email/password authentication, allowing personalized experiences and saved preferences.

• Frontend: React 18, TypeScript, Vite
• Styling: Tailwind CSS, shadcn/ui components
• Backend: Lovable Cloud (Supabase)
• Database: PostgreSQL with Row Level Security
• Authentication: Supabase Auth
• AI: Gemini & GPT models
• Routing: React Router v6
• State Management: TanStack Query

• Node.js (v18 or higher)
• npm or yarn
• Git

1. Clone the repository:

git clone <your-repo-url>
cd <project-name>

2. Install dependencies:

npm install

3. Set up environment variables: Create a .env file in the root directory with the following variables:

VITE_SUPABASE_URL=your_supabase_url
VITE_SUPABASE_PUBLISHABLE_KEY=your_supabase_anon_key
VITE_SUPABASE_PROJECT_ID=your_project_id

4. Start the development server:

npm run dev

The app will be available at http://localhost:8080

• npm run dev - Start development server
• npm run build - Build for production
• npm run preview - Preview production build locally
• npm run lint - Run ESLint

├── src/
│   ├── components/       # Reusable UI components
│   │   ├── ui/          # shadcn/ui components
│   │   └── auth/        # Authentication components
│   ├── pages/           # Page components
│   ├── integrations/    # Supabase integration
│   ├── hooks/           # Custom React hooks
│   └── lib/             # Utility functions
├── supabase/
│   └── functions/       # Edge functions
└── public/              # Static assets

The app uses the following main tables:

• chat_history - Stores conversation history
• suggestion_history - Tracks user decisions and suggestions
• image_analysis_history - Records fridge photo analyses
• user_preferences - Stores user preferences and settings

• universal-search - Handles AI-powered decision assistance
• generate-recipes - Creates recipe suggestions from ingredients
• generate-activities - Suggests activities based on context
• analyze-ingredients - Analyzes fridge photos for ingredients
• chat-modify - Manages conversational AI interactions

This app can be deployed to any static hosting platform:

• Vercel
• Netlify
• AWS Amplify
• GitHub Pages

Make sure to configure environment variables in your hosting platform.

This project is open source and available under the MIT License.