How to OCR a scanned PDF and get editable text

A scanned PDF is a set of photographs wrapped in a PDF container. There is no text inside, only pixels, so you cannot select a sentence, search for a name or copy a paragraph. OCR fixes that: it looks at the shapes on each page and rebuilds the letters.

The short answer

• Open Scanned PDF to text (OCR), load the PDF and choose the language of the document.
• Pages that already contain a text layer are read directly; only the image pages go through OCR.
• Download the result as a TXT file and proofread it before you send it anywhere.

The recognition runs in your browser. The PDF is never uploaded, which matters when the document is a contract, a medical report or an invoice.

Scanned or digital? Check first

Not every PDF needs OCR. Open the file and try to select a word with the mouse. If the text highlights normally, the document already has a text layer and you can extract it exactly with PDF to TXT. It is faster and avoids recognition errors.

If nothing can be selected, or the selection is a blue rectangle over the whole page, the page is an image. That is when OCR is the right tool.

Mixed documents exist too: a contract with digital terms and scanned signatures, or a report with a text cover and photographed annexes. A good OCR tool should treat each page on its own instead of forcing the whole file through recognition.

Step by step

• Load the PDF. Drop it into Scanned PDF to text (OCR) or click to browse. The file is opened locally.
• Choose the language. Pick the language of the text inside the document, not the language of your browser. English is added automatically so mixed pages still work.
• Start the conversion. Digital pages appear almost instantly. Scanned pages take a few seconds each, and the first run also downloads the language data.
• Review the text. Read it next to the original, starting with names, numbers and dates.
• Download the TXT. Use it in your editor, your notes app or any tool that accepts plain text.

How to improve accuracy

OCR is a best guess, and the quality of the guess depends on the quality of the image.

• Resolution. Around 300 dpi is the sweet spot for a scanned page. A phone photo taken close up often works better than a small screenshot.
• Contrast. Dark, clean text on a light background is ideal. Shadows, flash glare and coloured paper confuse the engine.
• Straight pages. Skewed scans break lines and merge words. Straighten the page before scanning when you can.
• Language. The wrong language model is the most common cause of broken accents and invented words. If a document mixes two languages, run the minority pages separately.
• Crop. Remove borders, hands, staples and logos so the engine does not try to read them.

What OCR still gets wrong

• Handwriting. Printed text is the target. Block letters sometimes work; cursive almost never does.
• Tables and columns. The engine reads text, not layout. Cells can lose their alignment and columns may be merged, so expect to rebuild tables by hand.
• Numbers and codes. The digit 0 and the letter O, or 1, l and I, are classic confusions. Check invoice numbers, IBANs and email addresses character by character.
• Small print and stylised fonts. Tiny footnotes, thin italics and decorative logos produce many errors.
• Punctuation. Quotes, dashes and currency symbols are small and easy to lose.

A spell checker will not catch a wrong digit. That is why the review step is not optional for anything financial or legal.

Privacy and limits

The conversion is local: the PDF is read, rendered and recognised on your device, and the text never leaves it. That is the main reason to prefer a browser tool over a service that asks you to upload the file.

Two practical limits are worth knowing. Very large scans use more memory, and recognition time grows with the number of pages. If a document has hundreds of pages, split it into parts and process them one at a time.

When OCR is not the answer

• The PDF already has text. Use PDF to TXT for an exact extraction.
• You need the original layout. OCR returns text, not design. For a faithful copy, keep the original PDF or convert it to Word with PDF to DOCX.
• You need a searchable PDF. OCR text can be pasted into a new document, but this tool returns TXT; it does not add an invisible text layer to the original file.
• The page is a photo, not a PDF. Use Image to Text (OCR) instead.

A quick checklist

• Select a word in the PDF first to confirm it is really scanned.
• Choose the language of the document.
• Compare names, amounts, dates and reference numbers with the original.
• Keep the original PDF until the text has been verified.
• Store the TXT where you can search it later.

Done this way, OCR turns an unreadable scan into text you can actually work with, without sending a confidential document to a stranger's server.