A web application for OCR processing using multiple providers. Upload documents (images or PDFs), select an OCR provider, and extract text.
- Multiple OCR Providers: Choose from Tesseract, Mistral OCR, or Google Cloud Vision
- Privacy Mode: Process documents entirely in your browser with Tesseract - data never leaves your device
- BYOK (Bring Your Own Key): Use your own API keys for cloud providers, stored securely in localStorage
- PDF Support: Full PDF processing on both server and client side
- Benchmarking: Compare up to 4 providers side-by-side, export results as JSON/CSV
| Provider | Type | Notes |
|---|---|---|
| Tesseract (Local) | Client-side | Privacy Mode - runs entirely in browser |
| Mistral OCR | Server/BYOK | Requires API key |
| Google Cloud Vision | Server/BYOK | Requires ADC or API key |
# macOS
brew install popplerbun installCreate a .env.local file with optional API keys:
MISTRAL_API_KEY=your_mistral_key# Optional if using BYOKFor Google Cloud Vision, use Application Default Credentials:
gcloud auth application-default loginOr set GOOGLE_APPLICATION_CREDENTIALS pointing to a service account JSON file.
bun dev # Start development server at localhost:3000
bun run build # Production build
bun run lint # Run ESLint
bun run test# Run test suite- Framework: Next.js 16 (App Router) with React 19 and TypeScript
- UI: shadcn/ui with Tailwind CSS 4
- OCR: tesseract.js, Mistral OCR API, Google Cloud Vision API
- PDF: PDF.js (client), pdftoppm (server)
app/ # Next.js App Router pages and API routes
components/ # React components
lib/ocr/ # OCR provider abstraction layer
providers/ # Server-side provider implementations
client/ # Client-side BYOK providers
tests/ # Test suite
MIT