From scanned PDF to searchable text in 30 seconds

Every organisation has a filing cabinet worth of scanned PDFs. Contracts scanned in 2004. Employee handbooks photographed in a hurry. Meeting minutes captured on a phone. None of them are searchable. All of them should be.

Here is the 30-second workflow to fix them.

Step one — Drop the file

Go to /tools/pdf-to-ocr. Drag your scanned PDF into the drop zone. The file is loaded into browser memory; it does not leave your device.

Step two — Pick a language

English is default. For Slovak, Czech, German, French, Spanish, Italian, Polish, Portuguese, Turkish or Ukrainian, choose the appropriate language from the dropdown. The relevant Tesseract model downloads and caches for future runs.

Step three — Wait a moment

A 20-page scan takes about 15 seconds on a modern laptop. Progress is shown per page. During processing your device fan may spin up briefly — that is CPU-bound OCR at work.

Step four — Download

You will receive a searchable PDF. The original scan is preserved as the visible image; an invisible text layer is added on top. In Adobe Reader, Preview, or any PDF viewer, Ctrl+F now works. Copy-paste works. Full-text search across your filesystem works.

Tips for better accuracy

• Scan at 300 DPI or higher. Below that, character recognition suffers noticeably.
• Prefer grayscale scans to full-colour — the LSTM engine ignores colour and colour scans just take longer to decode.
• For photographs of documents, hold the camera parallel to the page. Perspective correction helps but is not magic.