qpkit

Copy text from a scanned PDF or image

OCR Reader copies existing PDF text directly and uses Tesseract to recognize scanned pages and images. The output is editable text, not a reconstruction of the page layout.

Open OCR Reader

  1. Choose a readable source

    Use the original scan rather than a screenshot of its preview. Small text, tilted pages, shadows and heavy compression make recognition harder. PDF Toolkit can split out only the relevant pages of a long document.

  2. Set the language and extract

    Open the image or PDF and choose the recognition language. Let extraction finish, then review the editable text. Recognition runs in the browser.

  3. Correct and save the text

    Compare the result with the source before copying it or downloading TXT. If the output is poor, improve the source image instead of repeating recognition on the same unreadable pixels.

What to check

  • Dates, reference numbers and characters such as 0/O or 1/l.
  • Accented names, punctuation and line breaks.
  • The reading order of columns and tables.

What the download contains

TXT contains the extracted words. It does not add a searchable text layer to the original PDF or reproduce its formatting.

A scanned PDF or an existing text layer?

A PDF can contain page pictures, selectable text, or a mixture. OCR Reader reads an existing text layer directly; an image-only page needs recognition. If copying a page in another viewer already gives accurate words, OCR is unlikely to improve that text. For scans, the chosen language and the clarity of the source affect recognition.

Changing the recognition language

Choose the language actually used in the scan before correcting a long result. Changing it automatically rereads files that needed OCR; PDFs read entirely from an existing text layer are unchanged. Unedited OCR results can be replaced. If you have edited the text, it is protected and the new reading can be appended instead. Compare the passages before copying the result so you do not accidentally keep duplicate readings.

When extraction needs another step

A password-protected PDF is rejected
Open it with its password in an appropriate reader, or use an authorized unprotected copy. OCR Reader does not unlock it.
Letters are confused or missing
Use a sharper, upright source, the correct language and less compression.
A two-column page reads in the wrong order
Reorder the extracted text against the page. TXT does not preserve a table or a multi-column layout.

PDF and document tools