What is a scanned PDF, and why can't I select the text?

A 20-page PDF arrives by email from an office. You want to find the place where your name appears, so you press Ctrl+F and type it: "No results". Yet your name is right there on page two. You try to copy a paragraph instead; when you drag the mouse across it, the whole page turns blue, or nothing gets highlighted at all.

The explanation is simple: there is no text on that page, only a photo of text. When a paper document goes through a scanner or a phone camera and lands in a PDF, the page becomes a picture as far as the computer is concerned. You read letters; the computer only sees coloured dots. This guide shows how to recognise such a file, how to get the words back with free tools, and what this site can and cannot do for you here.

Three quick checks that give a scanned page away

Open the document in whatever you normally use: a web browser such as Chrome or Edge, Adobe Acrobat Reader, or the files app on your phone. If two of the three checks below come up positive, the file is scanned.

• Search: press Ctrl+F on Windows or Command+F on a Mac and type a word you can clearly see on the page. On a phone, tap the viewer's magnifying-glass icon. "No results" is the first sign.
• Select: hold the left mouse button and drag across a line. In a text PDF only that line is selected; on a scanned page either nothing is selected or the whole page lights up as a single box.
• Zoom: enlarge the page to 300-400 percent. Real text stays sharp at any size; text in a photo goes blurry at the edges and you start to see square pixels. A shadow along the paper edge, slightly slanted lines and the marks of punched holes are giveaways too.

What is actually inside the file?

A PDF exported from Word or an accounting program stores every letter separately: which character, in which font, at which point on the page. Search, selection and copying all rely on that information. A scanned PDF has no such layer; each page is a single image, and the PDF is merely the envelope that holds those images in order.

A side effect of this is file size. Ten pages stored as text often take a few hundred kilobytes; a photo of the same ten pages is measured in megabytes. If a short document is surprisingly large, that is another hint it was scanned.

A definite answer: try the PDF to text converter

If the three checks left you unsure, use this site's PDF to text converter. It reads the PDF's text layer; if that layer is empty it does not make anything up, it tells you there is no text layer. So if you get that warning, the document is definitely scanned. The file never leaves your computer while this happens.

Know one exception: some scanners and scanning apps place an invisible text layer behind the photo of the page. In such a file search works, the text can be selected, and the converter will pull out that hidden text. But that layer is the machine's guess: it can contain mistakes in numbers, names and accented letters. Never use an amount or an ID number you copied from it without comparing it with the document itself.

Getting the text back needs OCR, which this site does not have

The process that turns shapes in a picture into letters is called OCR, optical character recognition. To be clear: none of the tools on this site perform OCR. The PDF to text and PDF to Word converters stop and say so on a scanned page instead of giving you an empty file.

This site can still take you halfway. Most free OCR tools prefer images over PDFs, or only accept small files. Two of our tools turn a scanned page into a picture. The tool for extracting images from a PDF takes out the embedded scan exactly as stored, without recompressing it; if it meets a black-and-white compression type used by some scanners that it does not support, it tells you the image was skipped. In that case use the PDF to image converter to save the page as PNG or JPG at the resolution you choose; pick 300 DPI so the text stays crisp.

Free OCR: your phone and Google Drive

Most people already carry a text recognition tool in their pocket. Open the page you need on your phone and take a screenshot, or turn the page into an image with the tools above and send it to your phone. Then follow the steps for your phone.

• iPhone: open the image in the Photos app, tap the Live Text button in the lower-right corner or touch and hold a word, drag the grab points to widen the selection and tap Copy. The feature is available on iPhone XS and later with iOS 15 or later; if it is off, turn on Live Text under Settings > General > Language & Region.
• Android: open the image in the Google Photos app, tap More at the top and choose Lens, then select the words you want and copy them.
• Computer: upload the file to Google Drive, right-click it and choose Open with > Google Docs. Drive opens a new document with the text it recognised. According to Google's own help page, the file should be 2 MB or smaller, text should be at least 10 pixels high and the page should be the right way up; lists, tables and columns are unlikely to be detected. If your file is over 2 MB, cut it into a few pages at a time with the PDF splitter, or save the pages one by one as JPG.

Google Drive help: convert PDF and photo files to text →

Check the OCR result with your own eyes

Text recognition is a guess, and it goes wrong most often with look-alike shapes: zero and the letter O, the digit one and a lowercase l, rn read as m, accented letters losing their accents. Faint, tilted or shadowed scans make the errors more frequent. In an ordinary letter these are minor blemishes; in a bank account number, an ID number, an amount or a date they can have real consequences.

Make it a habit: after pasting the recognised text, put every string of digits side by side with the original document and compare. If reading a long text line by line is too much, at least check the names, the numbers and any tables. For a document full of tables, typing out just the rows you need is often quicker and safer than running the whole thing through OCR.

The opposite need: making your own PDF hard to copy

Sometimes you want exactly the reverse: when sharing a quote or a set of exam questions, you would rather the other side could not copy the text into their own document. The rasterize tool turns every page into an image, so your file behaves like a scanned document: search and selection no longer work.

Use it knowing that it is not a security measure. The phone features described in this guide can read the text on such a page in seconds. It only makes copying a nuisance; it does not protect information that has to stay confidential.

Frequently asked questions

Why doesn't Ctrl+F find anything in a scanned PDF?

Search works on the PDF's text layer. A scanned page has no such layer; the page is only a picture and the viewer cannot read the letters inside it. For search to work you first need a copy that has been through text recognition (OCR).

Can a scanned PDF be converted to Word?

Not directly; it needs OCR first. This site's PDF to Word converter does not do OCR and gives a warning on such a file. A free route is to upload the file to Google Drive, open it with Open with > Google Docs, and download the result as a Word file from the File menu. Always check the result by eye.

How do I copy text from a photo I took with my phone?

On an iPhone, open the image in Photos and tap the Live Text button, or touch and hold a word. On Android, open the image in Google Photos, tap More and choose Lens. Either way you can copy the selected text and paste it wherever you need it.

Why is a scanned PDF so large?

Because every page is stored as a photo, the file is many times bigger than a document stored as text. This site's PDF compressor can shrink it by recompressing the JPEG images of scanned pages; the pages stay photos, and the text does not become selectable.

The text in my PDF selects, but what I copy is gibberish. Is it scanned too?

Most likely, yes: during scanning an invisible text layer guessed by the machine was placed behind the page, and the guess was wrong. Sometimes the cause is instead a font whose character mapping is broken. Either way, the reliable route is to turn the page into an image, run it through text recognition again and compare the result with the document.

Last updated: October 12, 2026