Scan paper with your phone camera and keep it as clean, searchable PDFs. Works offline; scans never leave your device.
Android app (Capacitor) and Windows app (Electron) share one plain-JavaScript UI in app/.
- Live page detection in the camera, with auto capture when the page is held steady, torch, and Single / Batch / ID card modes
- Corner adjust with edge handles and a magnifier; "Whole photo" and re-detect
- Page detection finds the page by its straight edges (Hough lines,
detectByLinesinapp/js/imgproc.js) and prefers the outermost well-supported outline, so it does not stop at a header band, a table or a block of text - Perspective straightening (own homography + bilinear warp,
app/js/imgproc.js) - Filters: Original, Auto, Vivid, Clean, Grey, B & W, Ink saver, plus brightness and contrast
- Signatures: draw once, place and resize on any page
- Offline text recognition (Tesseract, English model bundled) with copy and share
- PDF export: A4 / Letter / Legal / Fit photo, three quality levels, optional searchable PDF (invisible text layer)
- Library: search by name or by text inside scans, sort, rename, reorder pages by drag, share as PDF / images / text
npm install
npm start # Windows/desktop app
npm test # unit tests (edge detection, warp, filters, PDF writer)
cd mobile && npm install && npm run sync # Android project in mobile/android
To measure page detection on synthetic photos of pages on desks:
python3 tools/bench/gen.py 77 /tmp/set # needs Pillow and numpy
node tools/bench/run.js app/js/imgproc.js 256 /tmp/set
CI (.github/workflows/build.yml) builds the Windows installer and a debug APK on every push.
Text recognition uses tesseract.js (Apache-2.0) and the Tesseract English model, vendored in app/vendor/tesseract.