Interior-Dost v2.9

AI-powered room redesign application with SDXL Turbo - fast, free, and excellent image-to-image support.

Overview

Interior-Dost is a production-ready MVP that accepts a single room image, analyzes it using vision AI, generates a redesigned version using SDXL Turbo, and suggests shoppable furniture categories. The core innovation is simplified, reliable image generation with strong geometry preservation for consistent, high-quality results.

Features

• 🖼️ Single room image upload (jpg/png) or camera capture
• 🤖 AI-powered room analysis (Gemini Vision or GPT-4 Vision)
• ✨ Detailed professional prompts ("Act as interior designer")
• 🎨 SDXL Turbo image generation - Fast, free, and excellent image-to-image support!
• 🔒 Perfect geometry preservation via image-to-image transformation
• 🏠 Realistic details: AC units, wiring, artwork, lighting fixtures
• 🛋️ Furniture category suggestions with shopping links
• 📱 Mobile camera support for instant photo capture
• 🎨 Premium interior design UI with warm neutral palette
• 🔄 Fallback modes for demo reliability
• ✍️ Personal prompt enhancement - write your own custom prompts

What's New in v2.9

🎯 Switched to SDXL Turbo for Better Geometry Preservation!

Focus on image-to-image quality - we've switched to SDXL Turbo which has superior image-to-image support:

• SDXL Turbo as Primary Model - Excellent image-to-image transformation!
• Better Geometry Preservation - Maintains room structure more accurately
• 90-Second Generation - Quick results without compromising quality
• 100% Free - No API costs for image generation
• Proven Quality - SDXL Turbo delivers consistent, professional results with better geometry lock

Result: Your room's windows, walls, and layout stay in place while furniture and decor are added! ✨

See WHATS-NEW.md for complete details.

Technology Stack

• Frontend: Next.js 14 (App Router), React 19, TypeScript
• Styling: TailwindCSS 4
• Vision AI: Google Gemini Vision API or OpenAI GPT-4 Vision API
• Image Generation: Pollinations.ai with SDXL Turbo (Fast, free, excellent image-to-image!)
• Testing: Jest, React Testing Library, fast-check (property-based testing)

Setup Instructions

Prerequisites

• Node.js 20+ and npm
• API key for Google Gemini Vision API (FREE tier available)
  • Get your key at: https://aistudio.google.com/app/apikey
• OR OpenAI API key (paid alternative for vision)

Note: SDXL Turbo is completely free - no API key needed for image generation!

Installation

1. Clone the repository:

git clone <repository-url>
cd interior-dost

2. Install dependencies:

npm install

3. Configure environment variables:

cp .env.example .env

4. Edit .env and add your Gemini API key:

# Vision AI (required)
GEMINI_API_KEY=your_gemini_api_key_here
VISION_PROVIDER=gemini

# Image Generation (no API key needed!)
# Pollinations.ai SDXL Turbo is used automatically - 100% FREE!

# Optional: Demo Mode
DEMO_MODE=false

5. Run the development server:

npm run dev

6. Open http://localhost:3000 in your browser.

Demo Flow

1. Upload or Capture Image: Select a room image (jpg or png, max 10MB) or take a photo with your mobile camera
2. Choose Style: Enter custom text or select a vibe preset (modern, traditional, minimalist, bohemian)
3. Generate: Click "Redesign My Room" to start the AI pipeline
4. View Results: See before/after comparison with geometry-preserved redesign, room analysis, and furniture suggestions
5. Shop: Click furniture categories to search on Amazon, Flipkart, or Pepperfry

What's New in v2.1

Critical Corrections & Polish

• ✅ Image Generation: Pollinations.ai SDXL Turbo as PRIMARY generator (FREE, no API key)
• ✅ Geometry Lock: Room structure preservation with explicit constraints
• ✅ Camera Capture: Real mobile camera support (not just claimed)
• ✅ Premium UI: Warm neutral color palette for interior design aesthetic
• ✅ Verified Navigation: Clean landing → design flow
• ✅ Clear Config: Simplified environment setup

See WHATS-NEW.md for full details.

Testing

Run tests:

npm test

Run tests in watch mode:

npm run test:watch

Generate coverage report:

npm run test:coverage

Known Limitations

• Single Image Only: Generates one redesigned image per request (no variations)
• Vision API Dependency: Requires active Gemini or OpenAI API key for room analysis
• No Persistence: No database - all data is session-based
• No Authentication: No user accounts or saved designs
• Image Generation: Uses Pollinations.ai SDXL Turbo (100% FREE, unlimited)
• Processing Time: Image generation typically takes 60-90 seconds
• Rate Limits: Subject to API provider rate limits (vision only - image generation is free)
• Indian Context: Optimized specifically for Indian home aesthetics

Project Structure

interior-dost/
├── app/                    # Next.js app router pages and API routes
├── components/             # React components
├── services/               # External API integrations
├── lib/                    # Utility functions and core logic
├── types/                  # TypeScript type definitions
├── __tests__/              # Test files
├── public/                 # Static assets
└── .env.example            # Environment variable template

API Costs (Approximate)

• SDXL Turbo Image Generation: 100% FREE (no API key needed)
• Gemini Vision: Free tier available, then ~$0.001 per image
• GPT-4 Vision: ~$0.01 per image

Budget accordingly for vision API usage. Image generation is completely free!

Contributing

This is a hackathon MVP. Contributions welcome but keep the scope focused on reliability and demo-readiness.

License

MIT License - see LICENSE file for details.

---

Built with ❤️ for Indian homes