# Why Can't I Copy Text From a PDF? Image-Only Scans

- Author: Schema Law (https://www.schemalaw.com/)
- Published: 2024-12-01T18:34:01Z
- Updated: 2026-10-11T19:22:20Z
- Category: Difficult scans
- Canonical: https://www.schemalaw.com/blog/why-cant-i-copy-text-from-a-pdf

> You can't copy text from a PDF when its pages are pictures with no text behind them. Here is how to tell, and how to get text you can select.

You can't copy text from a PDF when the file holds only pictures of pages. A scanner or fax machine saves each page as an image, so there are no letters behind what you see, and nothing to select, copy or search. Running the PDF through OCR adds the missing text.

That one cause explains most cases a law office meets, but not all of them. A PDF can also block copying on purpose, or carry text that copies as nonsense. The sections below show how to tell the three apart in about a minute, and what to do about each.

## What is an image-only PDF?

An image-only PDF is a file in which every page is a single picture. It looks like a document on screen, but it stores pixels, the way a photograph does. No characters are recorded, so a computer has nothing to match when you search and nothing to hand over when you copy.

Scanners, multifunction copiers and fax servers all produce this kind of file by default. So does a phone camera when you save a photo of a page as a PDF. Much of the paper that reaches a law firm arrives this way: productions from the other side, records from a provider, a signed agreement sent back by a client, an old order pulled from a paper file.

In a litigation file that picture is often a scanned order. You want selectable text so you can find a default judgment, a summary judgment, a writ, or a line on enforcement of judgment. Run recognition, then read the dollar figure and the date in the ordering paragraph against the page before you paste either one.

A PDF that was saved from a word processor is different. It stores the actual characters along with their fonts and positions, which is why you can select a sentence in it. That kind of file is often called a native or text-based PDF.

There is also a middle case, and it is the one you want. A searchable PDF keeps the picture of each page and adds an invisible layer of text placed exactly over the words in the picture. You still see the original scan, stamps and signatures included, but the file behaves like a text document. The guide to [how to make a scanned PDF searchable](https://www.schemalaw.com/blog/how-to-make-a-scanned-pdf-searchable) explains how that layer is built and how to check it.

## How can you tell whether a PDF has real text?

Try to select a single word. If your cursor draws a blue box across the page as if you were cropping a photo, or selects the whole page at once, the page is an image. If individual words highlight as you drag, the file has text.

Three more checks take a few seconds each and settle any doubt:

1. **Search for a word you can see.** Press Ctrl+F (Command+F on a Mac) and type a word that is plainly on the page. Zero results on a page full of that word means there is no text layer.
2. **Zoom in a long way.** At 400% or more, scanned letters turn soft and blocky at the edges, and you may see specks or a gray background. Real text stays sharp at any size.
3. **Look at the file size.** A ten-page contract saved from a word processor is usually well under a megabyte. The same ten pages as scans can run to several megabytes, because pictures take far more space than characters.

Check more than the first page. Files are often assembled from several sources, so a typed cover letter can sit in front of forty scanned pages. A search that finds a word on page one and nothing afterwards is the usual sign of a mixed file.

One trap is worth knowing about. Some scanners add a text layer on their own, and it can be poor. If you can select text but pasting it gives you something like "Tbe parlies agrcc", the file has a text layer that was recognized badly. The fix is to run recognition again with a better tool, and the post on [why OCR output is full of errors](https://www.schemalaw.com/blog/why-is-ocr-output-full-of-errors) covers how to judge the result.

## Other reasons copying fails: permissions and odd fonts

Two causes have nothing to do with scanning. The author may have set the file to forbid copying, or the file may use fonts that do not map back to ordinary characters. In both cases you can usually select words, which is how you tell them apart from a scan.

**Copy restrictions.** A PDF can carry permission settings that tell a reader program not to allow copying or printing. Open the document properties in your PDF reader and look at the security tab. If content copying is listed as not allowed, the block is deliberate. The honest route is to ask whoever sent the file for an unrestricted copy. That matters more in legal work than elsewhere, because the restriction may reflect a protective order or a licence.

**Fonts that copy as gibberish.** Some PDFs draw their letters with custom or subset fonts that do not record which character each shape stands for. The page reads normally, but pasted text comes out as boxes, symbols or the wrong letters. Court forms and older reports generated by case systems do this now and then. Since the page still looks right, the practical fix is the same as for a scan: treat the page as a picture and recognize the text from what is visible.

| What you see | Likely cause | What to do |
| --- | --- | --- |
| Cannot select any word; search finds nothing | Image-only scan | Run OCR on the file |
| Words select, but pasted text is full of wrong letters | Poor text layer from the scanner | Run OCR again with a better tool |
| Words select; copy is greyed out | Copy restriction set by the author | Ask the sender for an unrestricted copy |
| Words select; pasted text is boxes or symbols | Font without a character map | Run OCR on the page image |

## Getting selectable text out of an image-only file

Optical character recognition, or OCR, reads the picture of each page and works out which characters are in it. You upload the scan, the software recognizes the text, and you export the result in a format that holds real characters.

Which export you pick depends on what you need the text for:

- **To search or file the document as it is,** export a searchable PDF. The page images stay untouched and the text sits behind them.
- **To edit or reuse the language,** export Word. The post on [whether OCR keeps formatting](https://www.schemalaw.com/blog/does-ocr-keep-formatting) sets out what carries over.
- **To paste a passage into a brief or an email,** plain text is the quickest. The guide to [extracting text from a scanned PDF](https://www.schemalaw.com/blog/scanned-pdf-to-text) covers that route.

In Schema Law the steps are the same for each. Sign in, drop the PDF, PNG or JPEG onto the upload area, wait for the pages to be read, and choose an export format. The free plan covers one document a day of up to three pages and exports every format the paid plans do, which is enough to test the approach on a real page before you rely on it. The [pricing page](https://www.schemalaw.com/pricing) lists the page limits for larger files.

Whatever tool you use, compare the result with the original before you depend on it. Recognition is good on clean print and weaker on faxes, small type, stamps and handwriting. Read names, dates, dollar amounts and section numbers against the page image, because those are the characters where one wrong digit changes the meaning.

If the file is a client document, think about where you are sending it before you upload. Free converters differ a great deal in what they keep and what they do with it. The post on [whether it is safe to upload client documents to an OCR tool](https://www.schemalaw.com/blog/is-it-safe-to-upload-client-documents-to-an-ocr-tool) lists the questions to ask first.

## Try it on a real document

Upload a scanned PDF and export it as Word, a searchable PDF or plain text on the free plan. Your documents are never used to train an AI model. [Start free](https://www.schemalaw.com/api/auth/google)