Can't Copy Text From a PDF? Four Causes and How to Fix Them

You usually cannot copy text from a PDF for one of four reasons: the page is a scanned image, the creator blocked copying, the fonts cannot map shapes back to letters, or the text was converted to outlines or images. Each needs a different fix, from OCR to asking for an unrestricted copy.

The symptoms give you a clue. If nothing highlights when you drag across the page, there is probably no text at all. If text highlights but the copy option is greyed out, it is a permission. If it copies but pastes as gibberish or boxes, it is a font encoding problem. The sections below explain each cause, how to confirm it and what to do.

Cause 1: the page is a scanned image

A scanned page is a photograph of paper. It looks like text, but the PDF contains only an image, so there are no characters to select. Photos of documents taken on a phone and pages "printed" to PDF as images behave the same way.

How to check: try to select a single word. If the whole page highlights as one block, or nothing highlights, the page is an image. Searching for a word you can see on the page will also find nothing.

Fix: run optical character recognition (OCR), which reads the image and adds a text layer. Priya LifePDF's OCR PDF uses the open-source Tesseract engine in your browser, so the document is not uploaded. It keeps the scanned image and places the recognised words over it as invisible text, which you can then select, search and copy. Our guide on making a scanned PDF searchable walks through the settings.

OCR is not perfect, and the scan quality matters most. Tesseract's own documentation recommends images of at least 300 DPI, and that line segmentation gets significantly worse if a page is too skewed. Our tool renders each page before reading it; its highest setting, Ultra Resolution, renders at 216 pixels per inch, below that 300 DPI guideline, so small or faint print may still need a better scan. Always proofread copied numbers, names and units against the image.

Cause 2: copying is blocked by the document's permissions

A PDF can open without a password and still restrict what you do with it. The PDF standard, ISO 32000-1, defines permission flags that apply when a document is opened without its owner password (Table 22). Bit 5 controls copying or otherwise extracting text and graphics. Bit 10 separately controls extracting text and graphics in support of accessibility, so a file can block general copying while still allowing screen readers to read it.

How to check: text highlights normally, but Copy is greyed out or does nothing. Open the document's security properties in a full PDF reader: it will show content copying as not allowed.

Fix: if someone else created the file, ask them for an unrestricted copy. If it is your own document, or you have the right to change it, Remove Restrictions writes a copy with the encryption, and therefore the permission flags, removed. It cannot remove a password that is needed to open the file, and it invalidates any digital signature in the document, so keep the original. Our guide on removing a password from a PDF explains which of our tools fits which situation.

Only remove restrictions from documents you own or have permission to change. The restriction may be there because of a licence or a confidentiality agreement.

Cause 3: the text copies as gibberish

Sometimes you can select and copy the text, but it pastes as random symbols, empty boxes or letters that do not match. The cause is inside the fonts.

A PDF stores text as character codes that pick glyphs, the shapes drawn on the page, from a font. To copy, the reader has to turn those codes back into real characters, usually Unicode. Section 9.10 of ISO 32000-1 explains that this is easy when the font uses a standard encoding or standard character names. If it does not, the glyphs can still be shown, but the characters cannot be converted to Unicode without extra information, such as a ToUnicode map in the font. When a file lacks that information, the page looks perfect, but the reader has no reliable way to know which letter each shape is.

How to check: paste into a plain text editor. If the result is nonsense while the page looks normal, it is an encoding problem. Searching the PDF for a visible word will usually fail too.

Fix: the proper fix is to recreate the PDF from its source, with a current export method that embeds fonts properly. If you cannot, OCR is a workaround. Because the broken text is still in the file, first make an image-only copy with Rasterize PDF, choosing PDF as the output format and around 300 DPI, then run OCR on that copy. You lose the original text layer in the copy, so keep the original file for the record.

Cause 4: the text is no longer text

Design and print workflows can convert text to outlines, turning each letter into a vector shape. Some tools flatten pages to images to fix their appearance. Either way, the letters are now drawings: they print sharply, but there are no characters left to copy.

How to check: nothing selects, yet the text stays perfectly sharp when you zoom right in, which suggests vector outlines rather than a scan. If it turns blocky at high zoom, it is an image.

Fix: ask for the original file or a version exported with live text. Failing that, run OCR. Because OCR reads the rendered page, it works on outlined text as well as on images.

Other things to check

• The selection tool. Some readers have a hand tool for scrolling and a separate text selection tool. Make sure you are using the text one.
• Form fields and comments. Text inside a form field or a comment is not part of the page text and may need to be copied from the field or the comment itself.
• Mixed documents. A file assembled from typed pages and scans may have text on some pages and none on others. Check each problem page.
• A different reader. If one viewer will not copy but another will, the problem is the viewer, not the file.

Copying text from PDFs in regulated work

Being able to select and copy text matters well beyond convenience. Regulators reviewing submissions need it: the FDA's PDF Specifications ask that reviewers be able to select text and graphics, and the EMA's harmonised eCTD guidance requires that, where security settings cannot be removed, printing and content copying are allowed at a minimum.

When you copy data out of a PDF into another record, such as results from a certificate of analysis into a spreadsheet or a deviation report, treat OCR output as a transcription. Check it against the source image, and follow your SOPs on second-person verification where they apply. A misread decimal point or a confused "1" and "l" is easy to miss. OCR text added to a scanned record is also a change to that file, so keep the original scan as the record and use the OCR copy as a working copy, unless your procedures say otherwise.