Scanlite

Turn photos of documents into clean, searchable PDFs.

Load scanned pages or photos from your phone, let Scanlite straighten, crop, and clean them automatically, reorder as needed, and export a single PDF with an optional OCR text layer.

Getting Started

pip install scanlite
scanlite

Or download a native installer from Releases:

Native installers bundle Tesseract for OCR out of the box. When installing via pip, Tesseract must be available on your system PATH for the OCR export to work.

Features

Import

• PDF (rendered at 200 DPI), PNG, JPG, TIFF, BMP, WebP
• Multi-page PDFs split into individual pages automatically
• Drag-and-drop support (when tkdnd is available)

Processing

• Auto-crop: Otsu threshold + contour detection trims to the document edge

• Auto-perspective: Canny edge detection finds a 4-point quadrilateral and warps it flat

• Scan enhance: three modes that produce a clean, scanner-like look:

  Mode	What it does
  auto	Detects page brightness; picks B&W for white pages, CLAHE for darker ones
  bw	Adaptive Gaussian threshold for pure black-on-white text
  gray	CLAHE contrast enhancement for photos or diagrams

• Every operation works per-page or as a batch across all pages

• Reset any page to its original at any time

Reorder and Manage

• Move Up / Move Down (or arrow keys) to reorder
• Delete to remove a page
• Thumbnail panel shows the current page order at a glance

Export

• PDF: combines all pages into a single document
• PDF + OCR: Tesseract generates an invisible text layer per page, making the output searchable and copy-pasteable
• Keyboard shortcuts: Ctrl+S export, Ctrl+Shift+S export with OCR

Keyboard Shortcuts

How It Works

Scanlite's processing pipeline runs perspective correction first (to avoid cropping a skewed quad), then crop, then enhancement. Each step is independent and can be applied or skipped per page.

The perspective detector looks for the largest 4-sided contour in the edge map and computes a homography to warp it into a rectangle. If no quadrilateral is found, the image passes through unchanged.

The scan enhancer auto-detects whether a page is mostly white (median brightness > 170) and picks adaptive thresholding for text-heavy pages or CLAHE for photos and diagrams.

System Requirements

• Python 3.10+
• Tesseract OCR (bundled in native installers; for pip installs, install separately)

Support

"Software is like sex: it's better when it's free." -- Linus Torvalds

If this tool saved you a trip to the copy shop, buying me a coffee is a nice way to say thanks.

License

MIT (see LICENSE.md)

Citation

@software{scanlite,
  author = {Colling, Gilles},
  title = {Scanlite: Cross-Platform Document Scanner},
  year = {2026},
  url = {https://github.com/gcol33/scanlite}
}