Text recognition

Can't copy text from a PDF? Four causes and what helps

Can't copy text from a PDF? How to tell a scan, a broken font, copy protection and a viewer problem apart, and what fixes each one, without uploading the file.

Published 11 min read

When you can’t select or copy the text in a PDF, there is almost always one of four causes: the PDF is a scan, so the page is an image without text. The font is embedded in a way that turns pasted text into gibberish. The author has restricted copying. Or your PDF viewer is using the wrong tool. The first two are solved by text recognition (OCR) on your PC, the third only by the author, the fourth by a click in the viewer.

This guide shows how to find the cause in a minute, what helps in each case, and how to get the text out of a scanned PDF without uploading the document anywhere.

Find the cause in a minute

What happensCause
Dragging with the mouse selects nothing, and searching the PDF finds no wordsScan: the page is an image
Selecting works, but the pasted text shows boxes, question marks or wrong lettersFont: the font is embedded without a readable character mapping
Selecting works, but “Copy” is grayed out or does nothingCopy protection: the author set a permissions password
Nothing works in the browser but it does in a PDF app, or the other way roundViewer: wrong tool or wrong mode

A quick hint for a scan: press Ctrl+F and search for a word you can see on the page. If the search finds nothing and nothing on the page can be selected, the page is most likely an image. If you can select text but the search doesn’t find it, the font is the more likely cause (cause 2). Some PDFs mix scanned pages and real text pages, so check the page you need.

Cause 1: the PDF is a scan

To your computer, a scanned document is an image. The text in it isn’t type but a pattern of pixels. You can’t select it, search it or copy it. Adobe says the same: when you scan a paper document to PDF, the resulting file contains only image data, not searchable text.

What helps is text recognition (OCR), which turns the image back into text. There are three ways to do that:

  • Snipping Tool in Windows 11. After a capture with Win+Shift+S, the “Text actions” button extracts the text from the image; you can then select and copy it. According to Microsoft, text recognition runs locally on your device. Microsoft describes the feature only for Windows 11, not for Windows 10, and doesn’t say how a poor original is prepared before recognition.
  • PowerToys Text Extractor. Win+Shift+T, drag a region, and the text is on the clipboard. It only recognizes languages that have the Windows OCR language pack installed, and Microsoft itself says the output needs a quick proofread. Microsoft now recommends the Snipping Tool instead.
  • Acrobat. Text recognition (“Scan & OCR”) is part of Acrobat, not of the free Acrobat Reader. Reader’s help only covers the Select Tool and the Snapshot tool, which copies an area as an image.

Online OCR services are the fourth way. They require you to upload the document. For a scanned contract, an HR form or a medical letter, that’s not a small thing.

Copy text from a scanned PDF: with Celerando on your PC

Text recognition in Celerando works on what’s on the screen. You open the PDF as usual, drag a frame over the page, and the recognized text is on your clipboard. Recognition runs on your computer, with Tesseract 5. Neither the image nor the recognized text leaves it. Celerando runs on Windows 10 and 11.

Edition limits: text recognition is included from Celerando Pro (€36 one-time, no subscription). You can try all features free for 14 days; afterwards, Celerando keeps running as Free, without text recognition. All editions are on Editions and pricing.

Step by step

  1. Install Celerando. The installer is signed and needs no admin rights or account.
  2. Open the PDF and zoom in. The larger the type on screen, the better the result. In the Screenshot and Scan / document profiles, Celerando enlarges the image 1.5× before recognition. In the Pro profile you can set the factor up to 3×. A clearly readable original is still the safest route.
  3. Press Ctrl+Alt+T and drag a frame over the text. The frame can span several monitors. You can change the hotkey in Configuration.
  4. Ctrl+V in the target app: Word, email, browser, editor. The text is on the clipboard right away, without saving or exporting anything.
  5. For poor scans, choose the Scan / document profile. You’ll find it under Configuration → OCR → OCR profile. Celerando prepares the image before recognition: scaling, greyscale, contrast, binarisation, sharpening and noise reduction, bundled into the Screenshot, Photo / camera and Scan / document profiles. Scan / document is meant for PDF pages and paper.

For longer documents, repeat steps 3 and 4 page by page. A German-English contract is recognized in one pass; the default selection is German + English, and twelve more languages can be added.

Cause 2: selecting works, but the pasted text is gibberish

Here the PDF does contain text, but the font is embedded without a usable mapping to normal characters. On screen everything looks right, because the viewer only draws the letter shapes. When you copy, the information about which letter is behind which shape is missing, and you get boxes, question marks or shifted letters. This happens with some export tools, with fonts embedded only as outlines, and with print-shop files.

What helps: the same text recognition as for a scan. It reads what’s on the screen, and with these PDFs that is correct. So zoom in, Ctrl+Alt+T, drag the frame, Ctrl+V. The Screenshot profile is meant for screenshots and on-screen text.

If the document is yours, export it again, with the fonts as text rather than outlines. Fully embedding the fonts doesn’t guarantee copying on its own; what matters is that the export writes the mapping from letter shapes to characters. Check the result by copying and pasting a paragraph.

Cause 3: the author has restricted copying

PDFs can carry a permissions password. According to Adobe, it lets the author restrict certain features, such as printing, editing, copying content and adding comments. You can open and read the document as usual; only “Copy” stays grayed out or does nothing.

How to check in Acrobat Reader: right-click the document, choose “Document Properties”, then the “Security” tab. The Document Restrictions Summary there shows, among other things, whether content copying is allowed.

What helps: ask the author or the organization the PDF came from for a version without the restriction, or for permission. If it’s your own document, enter your permissions password. This guide doesn’t explain how to bypass the restriction: it is the author’s decision, and laws such as 17 U.S.C. § 1201 in the US and § 95a UrhG in Germany protect technical protection measures. Celerando’s text recognition is meant for scans, photos and screenshots, not for that.

Cause 4: it’s the viewer

Sometimes the PDF is fine, and only the tool is wrong:

  • Wrong tool in Acrobat Reader. You select text there with the Select Tool: right-click the document, choose “Select Tool”, drag over the text, right-click, “Copy”. If another tool is active, such as the one for moving the page, the mouse selects nothing.
  • PDF in the browser. The PDF viewers built into browsers can do less than a PDF app. If you can’t select anything in the browser, download the file and open it in a PDF app. Or drag a frame over the page in the browser with Celerando; text recognition doesn’t care which app shows the PDF.
  • You only want an image. For a table or a chart you want to pass on unchanged, Acrobat Reader has the Snapshot tool (“Edit” › “More” › “Take A Snapshot”). It copies the area to the clipboard as an image, not as text.
  • Text in form fields or stamps. Form fields, stamps and comments aren’t part of the page text and are skipped when you select the page. A frame with text recognition picks up whatever is on the screen.

When the recognized text has errors

Any text recognition produces poor text from a poor original. Here’s what you can do when the result isn’t right:

  • Zoom in. Increase the zoom in the PDF viewer until the letters on screen are clear, and drag the frame again. Tiny type is the most common cause of garbled letters.
  • The right profile. Scan / document for scanned pages and paper, Photo / camera for photographed pages with uneven light, Screenshot for screenshots and code. Binarisation helps a clean scan and destroys half the letters in a photo with shadows; that’s why there are three presets instead of one “enhance” button.
  • The right languages. The default is German + English. A French scan needs French in the selection; eight languages come with the installation, six more (Danish, Swedish, Finnish, Norwegian, Greek, Russian) you download under Configuration → OCR → Text recognition languages, one model of a few megabytes per language.
  • Adjust manually. In the Pro profile you control scaling, greyscale, contrast, binarisation, sharpening and noise reduction yourself, including the thresholds.
  • Accept the limits. A page a human can barely decipher won’t become clean text here either. Then the only option is to check the doubtful passages by hand.

Celerando and the Snipping Tool compared

For a clean screenshot in your system language, the built-in tool in Windows 11 is enough. The differences show up with scans, photos and small type.

CelerandoSnipping Tool “Text actions”
Where recognition runsOn your PCOn your PC
Image preparationSix stages, three profiles, upscaling 1.5× (up to 3× in the Pro profile)Not described by Microsoft
Languages14, several at onceNot stated on Microsoft’s help page
Multiple monitorsYes, also across monitorsNot stated on Microsoft’s help page
Windows10 and 11Described for Windows 11
PricePart of Pro, €36 one-time, 14-day trialIncluded in Windows 11

As of October 2026. PowerToys Text Extractor also works locally, with Win+Shift+T, but only with the OCR language packs installed in Windows; Microsoft now recommends the Snipping Tool instead.

Frequently asked questions

Why can’t I copy text from some PDFs?

Because the PDF either contains no text at all (a scan is an image), the font is embedded without a readable character mapping, the author restricted copying with a permissions password, or the wrong tool is active in the viewer. The table above shows how to tell which one it is.

How do I copy text from a scanned PDF into Word?

With text recognition. On Windows 11, the Snipping Tool with “Text actions” works for a clean original. With Celerando, you drag a frame over the page (Ctrl+Alt+T) and paste the text into Word with Ctrl+V; for poor scans, the Scan / document profile helps. Text recognition is included from Pro.

How do I copy text from a protected PDF?

You don’t, as long as the restriction applies. In Acrobat Reader, check under “Document Properties” › “Security” whether content copying is allowed, and ask the author for a version without the restriction or for permission. You unlock your own document with your permissions password.

Is there a free way?

On Windows 11, yes, with the Snipping Tool; PowerToys is free as well. In Celerando, text recognition is included from Pro (€36 one-time) and free to try for 14 days; the free edition doesn’t include it.

Do I have to upload the PDF anywhere?

No. Celerando recognizes text on your computer; with the eight bundled languages, text recognition needs no connection at all, which you can try in flight mode. Celerando only connects to the internet for the downloads you choose (models, extra languages, the optional CUDA runtime), the update check, license activation including the query of seats in use, a one-time installation ping on first launch, and addresses you enter yourself (web app tiles load the site’s icon, the “Connections” tile checks your servers). Your PDF is never part of this.

Does this work with a photo of the document from my phone?

Yes. Send the photo from your phone to your PC, open it on screen, and drag the text recognition frame over it. The Photo / camera profile is meant for photographed pages with uneven light.

Sources