Textract is a desktop OCR app for anyone digitizing contracts, IDs, receipts, and scans — documents you least want uploaded to someone else’s server. Adobe Acrobat Pro bundles OCR into a $19.99/mo subscription ($240/yr, forever); ABBYY FineReader is $99/yr; online OCR sites make you upload sensitive paperwork. Textract is $19 once and runs 100% on your machine — one month of Acrobat costs more than Textract, forever.

The entire pipeline is local: tesseract.js pointed at bundled worker/core/language files, never a CDN. The only network call the app can make is an optional extra-language download.

• Drag-drop OCR for images (PNG/JPG/WEBP/BMP/TIFF) and PDFs, per-page text with confidence scores
• Searchable-PDF export: invisible text layer positioned from OCR word boxes — Ctrl-F-able in any reader
• Batch folder mode (recursive optional): a .txt per file plus a files/pages/words summary
• Screenshot-to-text global hotkey (Ctrl+Shift+T) with region capture, even minimized to tray — auto-copies to clipboard
• Multi-language: English bundled, more download on demand
• Preprocessing pass (grayscale + adaptive threshold) rescues low-contrast scans
• Local history — re-copy past extractions without re-OCR
• Zero telemetry, no account, no uploads

Runs as a Windows desktop app.

What’s included:

• Windows installer download (build via npm run dist)
• Full source access — https://github.com/bensblueprints/textract-ocr (MIT)