Scribe - OCR Document Processing App

A SvelteKit 5 application for intelligent document processing using hybrid OCR: Mindee (official documents) and Mistral AI (creative content).

Features

• Hybrid OCR System: Automatic routing between Mindee (invoices, official docs) and Mistral AI (recipes, notes)
• Mobile-First: Optimized for document scanning with camera capture
• Validation Interface: Two-column layout for reviewing and editing extracted data
• Local Database: SQLite with Drizzle ORM for fast local development
• Modern UI: Built with Skeleton UI and Tailwind CSS
• Type-Safe: Full TypeScript with strict mode

Tech Stack

• Framework: SvelteKit 5 with runes
• UI Library: Skeleton UI + Tailwind CSS
• Database: SQLite with Drizzle ORM
• OCR Providers:
  • Mindee API (official documents, invoices)
  • Mistral AI (creative content, recipes, notes)
• Testing: Vitest (unit) + Playwright (E2E)
• Deployment: Cloudflare Workers

Getting Started

Prerequisites

• Node.js 18+ or Bun
• Mindee API Key
• Mistral API Key

Installation

# Clone the repository
git clone https://github.com/your-username/scribe.lelab.git
cd scribe.lelab

# Install dependencies
bun install

# Create .env file
cp .env.example .env
# Edit .env with your API keys

Environment Variables

MINDEE_API_KEY=your_mindee_api_key
MISTRAL_API_KEY=your_mistral_api_key
DATABASE_URL=.data/db.sqlite  # Optional, defaults to local file

Development

# Start development server
bun dev

# Generate database migrations
bun run db:generate

# Run migrations
bun run db:migrate

# Type check
bun run check

# Format code
bun run format

Project Structure

src/
├── routes/              # SvelteKit pages
│   ├── +page.svelte       # Upload interface with mobile-first design
│   ├── +page.server.ts    # Server actions (upload, OCR, save)
│   ├── validate/         # Validation page
│   │   ├── +page.svelte      # Two-column review interface
│   │   └── +page.server.ts   # Save document action
├── lib/
│   ├── db/
│   │   ├── schema.ts       # Database schema (SQLite)
│   │   └── index.ts       # Drizzle client
│   ├── supabase-storage.ts # Local file storage
│   ├── ocr-utils.ts      # OCR utility functions (normalize, parse, extract)
│   └── ocr-utils.spec.ts # Unit tests for OCR utilities
e2e/                   # E2E tests with Playwright
├── migrations/            # Database migrations
└── ai/                  # AI context and documentation

How It Works

1. Upload: User captures or selects a document image
2. OCR Processing:
   • Official documents → Mindee API
   • Creative content → Mistral AI
3. Validation: Two-column interface to review extracted data
4. Storage: Validated data saved to SQLite database

Contributing

This is an open-source project! Contributions are welcome.

1. Fork the repository
2. Create a feature branch (git checkout -b feat/my-feature)
3. Commit your changes (git commit -m 'feat: add my feature')
4. Push to the branch (git push origin feat/my-feature)
5. Open a Pull Request

License

MIT License - see LICENSE file for details

Built With

• SvelteKit - Web framework
• Skeleton UI - Design system
• Drizzle ORM - Database toolkit
• Mindee - OCR API for documents
• Mistral AI - AI for creative content