Scan to Markdown

Browser-based batch OCR tool — convert scanned PDFs and images to Markdown.

Powered by PaddleOCR.js, runs entirely in the browser. No server, no upload, no cost.

Features

• Batch upload — PDF files and images (PNG, JPG, TIFF, WebP)
• PDF rendering — extract pages as images via pdf.js
• Page management — thumbnails, drag-to-reorder, include/exclude toggles
• Full OCR control — detection thresholds, recognition params, language, model selection, custom pipeline config
• Post-OCR editing — edit recognized text, toggle text regions, draw custom regions
• Markdown export — per-page files or single merged document
• ZIP download — packaged with fflate
• 100% browser — no data leaves your machine

Quick Start

npm install
npm run dev

Open http://localhost:5173. The app requires Cross-Origin-Isolation headers (set automatically by Vite config) for threaded WASM and WebGPU support.

Supported Languages

• Chinese (Simplified): ch
• Chinese (Traditional): chinese_cht
• English: en
• Japanese: japan

Build

npm run build

Output goes to dist/. Serve with any static file server — make sure to set COOP/COEP headers:

Cross-Origin-Opener-Policy: same-origin
Cross-Origin-Embedder-Policy: require-corp

Tech Stack

• PaddleOCR.js — ONNX-based OCR in the browser
• pdf.js — PDF rendering
• fflate — ZIP compression
• Vite — build tooling
• Vanilla TypeScript — zero framework overhead

License

MIT