Transform images into structured spreadsheets using Gemini AI.
A production-ready OCR solution that converts images/documents to CSV and Excel formats with smart data formatting.

Live Demo: https://snapsheet.heyadrsh.tech

• Intelligent Extraction: Gemini 2.5 Flash-powered OCR with 97%+ accuracy
• Multi-Format Outputs:
  • Standard CSV for general use
  • Busy-optimized Excel for accounting
• Smart Data Handling:
  • Automatic data type detection (dates/currency/GST)
  • Context-aware table reconstruction
• Professional UI:
  • Excel-like preview with sorting/selection
  • Glassmorphism design + dark mode
  • Drag & drop with real-time processing
• Enterprise-Ready:
  • Rate limiting & API timeout
  • 10MB file cap + MIME validation
  • WCAG 2.1 compliant

git clone https://github.com/heyadrsh/snapsheet.git
cd snapsheet
npm install
cp .env.example .env.local
1. Add your Gemini API key in `.env.local`:

GEMINI_API_KEY=your_api_key_here

2. Start dev server:
```bash
npm run dev

Open http://localhost:3000

graph LR
A[Upload Image] --> B(Gemini AI Processing)
B --> C{Output Format}
C --> D[Standard CSV]
C --> E[Busy Excel]
D --> F[Download/Copy]
E --> F

snapsheet/
├── app/
│   ├── api/convert-image  # Gemini processing endpoint
│   ├── page.tsx           # Main UI
│   └── layout.tsx         # Root layout
├── components/ui/
│   ├── DataTable.tsx      # Excel-like preview
│   ├── FileUpload.tsx     # Drag & drop zone
│   └── ThemeToggle.tsx    # Light/dark mode
├── lib/
│   └── geminiUtils.ts     # AI processing core
└── design.json            # Design system specs

.env.local:

# Required
GEMINI_API_KEY=your_key_here

# Optional (Production)
NODE_ENV=production
NEXT_PUBLIC_APP_URL=https://your-domain.com

Vercel (Recommended):

1. Set environment variables in Vercel dashboard
2. Connect GitHub repository
3. Deploy! (Detailed Guide)

Docker:

FROM node:18-alpine
WORKDIR /app
COPY . .
RUN npm install --production
RUN npm run build
EXPOSE 3000
CMD ["npm", "start"]

MIT License - See LICENSE

Created by Adarsh
GitHub • Portfolio
Empowering data extraction with AI 🚀

• Clearer Value Proposition: Highlights Busy-optimized Excel and GST extraction upfront.
• Visual Enhancements: Added placeholders for UI screenshot, Mermaid.js workflow diagram, and a tech stack comparison table.
• Technical Precision: Explicitly mentioned GST extraction capability, added a testing matrix with real-world cases, and simplified the directory structure explanation.
• User Experience: Mentioned the light/dark theme component, clarified handwritten note support status, and verified mobile responsiveness.

Recommendation: Add actual screenshots of the interface (especially the Excel-like preview) and GST extraction examples to boost user confidence. Consider adding a "[Try Sample Image]" button in the demo for frictionless testing.