Why you cannot select or copy text in a PDF

Run the file through PDF to Text. Empty pages in the TXT mean a scan or outlined text, which needs OCR. A copy restriction without an open password does not block extraction. Garbled text needs the pages turned into images before OCR.

The quickest way to find the cause is to run the file through PDF to Text and look at the TXT it produces. The tool takes the PDF's text layer as it is: if there is no layer, the file comes out almost empty, and if the layer is damaged, you see the same strange characters you get when copying. That result tells you more than a greyed-out button in your viewer.

There are usually four causes: the page is a scan or a photo, the text was converted to outlines on export, the author restricted copying in the security settings, or the link between how letters look and what they mean is broken. Below is how to recognise each one and what to do about it.

Start with one line

Open the PDF, try to select an ordinary sentence and paste it into a plain text editor. Search for a word that is clearly visible on the page. Repeat on another page: one document can mix scanned pages with pages that hold real text.

Note what happens. Does nothing get selected, does the whole page get selected as one block, do random letters come out, or does the viewer say outright that copying is not allowed? These observations are more useful than the file size. A large PDF is not necessarily a scan, and a small one does not necessarily contain text. Also check that the viewer's text selection tool is active, not the hand tool for moving the page.

What the TXT from PDF to Text tells you

Upload the file and run the conversion with the default settings. In the TXT, every page starts with a marker such as "--- Page 3 ---". Under the marker you will see one of three things.

• "(empty page)". That page has no text layer: it is an image or outlined text. If every page looks like this, the job also shows the warning "This PDF has no text layer".
• Normal text. The layer is fine, and the problem is the viewer or the security restrictions described below.
• Random letters, boxes or the wrong alphabet. A layer exists, but it is damaged.

An empty page in the TXT does not mean the job failed: the tool is simply telling you there was nothing to extract.

The page is a scan or a photo

The letters may be part of a picture. The viewer draws them the same way it draws table lines, and extraction cannot turn an image into characters. This is a job for OCR PDF: it recognises the text and adds an invisible layer under the image, after which search and copying work.

The "OCR language" field offers three choices: "Auto", "English" and "Russian". "Auto" reads English and Russian together, which suits Russian and mixed documents. If the document has no Russian text, choose "English": in "Auto" mode, Latin letters can be recognised as look-alike Cyrillic ones. No other languages are available right now: German, French or Spanish text is recognised without its own language model, so letters with accents (ä, é, ñ) may be lost or replaced. For a tilted scan, turn on "Deskew before OCR", and in "OCR quality mode" choose "Accurate" when precision matters more than speed.

Recognition quality depends on the source. Small print, skew and noise cause errors, especially in numbers. On an invoice for 12,580, check the amount, the number and the date separately: one correctly recognised surname is not enough to trust the whole page. If a sharper scan exists, start from that. The settings are covered in more detail in How to recognize text in a scanned PDF using OCR.

The text was converted to outlines

Sometimes a page looks like an ordinary typeset document, yet no letter can be selected and the TXT shows "(empty page)". A common reason is that the text was converted to outlines (curves) when exporting from a design program or preparing for print. Each letter became a vector shape, like a line or a logo, and the file no longer holds characters. If the letters stay sharp when you zoom in a lot, it is not a scan, but for copying it makes no difference: the page behaves like a picture.

OCR helps here as well. If you open the file in Edit PDF and click "Recognise text (OCR)", the panel shows the state of each page: "has text", "scan" or "drawing". The "drawing" label marks a page with no text but a lot of vector graphics, which is typical of outlined text. The "Pages without text" option selects both scans and such pages.

Copying is restricted by the security settings

A PDF can open without a password and still forbid copying: the author set restrictions with an owner password. The viewer then greys out "Copy" or shows a message about permissions. This is not the same as an open password, without which the file shows no pages at all.

If the file opens without a password, PDF to Text extracts the text straight away; you do not need to remove anything first. If you need to copy directly in the viewer, Unlock PDF removes such restrictions: leave the "Current password" field empty, as no owner password is needed. Do this with your own documents or with files you are allowed to use. Restrictions can be part of an agreement with the author, and removing them does not give you the right to reuse the content.

If the file asks for a password to open, you need to know it: enter it in "Current password". The tool does not guess passwords. Without the password, ask the sender for an accessible version. Protection is covered in How to Remove a PDF Password.

Copied text turns into gibberish

A PDF stores how a letter looks separately from what it means. If the font's mapping table is incomplete or damaged, the page reads normally, but nonsense lands on the clipboard. The usual culprit is the font or the way the source document was exported. Renaming the file to .txt fixes nothing.

The best route is a fresh export from the original program with fonts embedded. If you do not have the source, watch out for a trap: OCR PDF skips pages that already have a text layer of at least 24 non-space characters, even if that layer is garbage. The job finishes, but those pages stay as they were.

The workaround is to turn the pages into images first. Run the file through PDF to JPG: you get a ZIP archive with one image per page. Upload the whole archive or the unpacked images to JPG to PDF and check the page order. The new PDF has no text layer, so OCR PDF will recognise every page. Links and bookmarks from the original are lost and the file may get bigger, so keep the original alongside. In one run, JPG to PDF accepts a limited number of images, and images inside a ZIP count too (the limit is shown on the tool page). Process a long document in parts and join the resulting PDFs with Merge PDF. The workaround takes at least three operations from your daily limit, and more if you work in parts.

Check the whole result

Once fixed, open the first, a middle and the last page. Search for a word, copy a sentence and compare key values with the image. Check tables separately: having text does not mean the cells are in the right order in the TXT.

If you need all of the text rather than a single quote, move on to extraction: choosing a mode and setting page breaks is explained in how to extract text from a PDF and save it as TXT. Keep the original PDF next to the result and note the places where recognition or line order needed manual fixes.