Scanned PDF OCR

OCR scanned PDFs in the browser

When a PDF contains only scanned images, CrabPDF can run OCR, create editable text overlays, and let you correct recognised words locally without uploading the file.

How to OCR a scanned PDF

Upload a scanned PDF that contains image-only pages.

Click Run OCR in the right sidebar.

Wait for the browser to process each page with Tesseract.js.

Review the detected OCR words overlaid on the document.

Double-click any word to correct it inline.

Open the OCR Fix sidebar to review all low-confidence words at once.

What to expect

OCR creates editable overlays from scanned text regions.

Edits on scanned pages cover the original image area and draw new text.

OCR quality depends on scan resolution, language, and document noise.

Low-confidence words are flagged automatically for review.

Private by design

OCR runs entirely in your browser using Tesseract.js. No page images are uploaded to any server.

CrabPDF is experimental. Always verify important documents after OCR and editing.