How to get editable text out of a scanned PDF

You try to copy a paragraph from a PDF and nothing highlights. You press Ctrl+F and the search finds nothing, even though the word is right there on the page. The PDF is a scan: each page is a photograph of paper, with no text inside it. To edit, search or quote it, the text has to be read from the pictures with optical character recognition, or OCR.

Check whether the PDF has text

Try selecting a line with the mouse. If whole pages highlight as a block, or nothing highlights, it is a scan. For a definite answer, open PDF to Text: it reports which pages contained no text. If every page has text already, you do not need OCR; copy it directly or use PDF to Word.

Run OCR on the pages

1. Turn the pages into images with PDF to PNG. PNG keeps every pixel, which helps recognition more than JPG.
2. Open Image to Text and add a page image.
3. Choose the language of the text. Recognition is much better with the right language set.
4. Copy the result into your document and repeat for the other pages.

Both steps run in your browser. That matters for scanned contracts, payslips and medical records, which should not be uploaded to a stranger's server just to read them.

Get better results

• Straighten pages first. Sideways or upside-down pages are read poorly. Fix them with Rotate PDF before exporting.
• Resolution matters. Small print scanned at low resolution blurs into shapes. If you can rescan, 300 dpi is the usual recommendation for OCR.
• Contrast matters. Faint photocopies, coloured paper and stamps over text all reduce accuracy.
• Simple layouts work best. Single columns come out cleanly. Multi-column pages, tables and forms may be read across the columns, so check the order.

Check the text carefully

OCR is very good on clean print but never perfect. The errors are predictable, so look for them:

• Look-alike characters: 0 and O, 1, l and I, 5 and S, rn and m.
• Numbers in amounts, dates and account numbers, where one wrong digit matters most.
• Words split by line-end hyphens.
• Missing accents and symbols.

If the extracted text has broken line breaks and odd spacing, the Text Cleaner joins lines and fixes spacing in one pass. The guide to fixing text copied from a PDF covers the same clean-up for text PDFs.

Make a searchable copy for later

Once the text is out, you can keep it in two forms: an editable document for quoting and reusing, and the original scan as the record of what the paper said. Store them side by side with matching names, such as lease-2026-scan.pdf and lease-2026-text.docx. Your computer's file search will then find the lease by any word inside it, which the scan alone could never do.

Compressed PDFs lose their text too

A PDF that was originally typed can still end up as pictures. Heavy compression often redraws pages as images, and printing to a scanner-style “image PDF” does the same. If an older copy of the file exists, use that instead of running OCR on the compressed one. The guide to copying text from an image covers OCR on screenshots and photos.