How to turn a scanned PDF into an editable Word document

By Uxair Ijaz · 2026-10-22 · 2 min read

Some PDFs are really photographs of paper: a scanned contract, an old report, a phone snap saved as PDF. You cannot select their text because there is no text, only pixels. OCR (optical character recognition) reads the letters in the image and turns them back into real text you can edit.

Is my PDF scanned?

Open it in the PDF viewer and try to select a sentence. If you can highlight words, the PDF has a text layer and converts directly. If the whole page selects like one picture, it is scanned and needs OCR.

Converting with OCR

Open PDF to Word.
Leave OCR on "OCR scanned pages" (the default). Pages that already have text are read directly; only pages without text go through OCR.
Choose the language of the text, for example English, Spanish or Arabic. The right language makes a big difference to accuracy.
Drop your PDF. A progress bar shows each page being read.
Check the preview, then download the .docx and edit it in Word, Google Docs or LibreOffice.

The first time you use a language, its recognition data (a few megabytes) downloads once and is then cached. Your PDF itself never leaves your computer: the OCR runs in your browser.

Getting the best results

Resolution: scans at 300 dpi read far better than 150 dpi faxes.
Straight pages: rotate sideways scans first with PDF tools.
Clean contrast: dark text on a white page. Shadows, stamps and handwriting reduce accuracy.
One language per run: if a document mixes languages, convert it with the main one.
Proofread names, numbers and amounts. OCR is good, but a 0 and an O can still swap.

What comes across

With Exact layout (the default), each Word page looks like the scanned page: the recognised text becomes editable text where it was printed, and the rest of the scan (stamps, signatures, table lines) stays behind it as an image. Choose Flowing text when you mainly want to edit: it rebuilds paragraphs and headings without the page image.

Choosing "OCR every page"

Some PDFs have a broken or garbled text layer (copying gives nonsense characters). Switch OCR to "OCR every page" to ignore the text layer and read every page from its image instead.