How to See the Text Layer in a PDF Behind a Scanned Page

To see the text layer in a PDF, select everything on a page, copy it, and paste it into a plain text editor. What lands in the editor is the hidden text an OCR tool stored behind the scanned picture, character for character. To read the layer for a whole file, export the PDF to plain text.

The layer is invisible on purpose, so the page keeps looking like the scan. That also means nobody reads it. Search and copy both run on those hidden characters, and the only way to know what they say is to bring them out where you can look at them.

Paste a caption page and a later page into a text editor. Find the plaintiff, the defendant, the docket number, and a Bates number in the paste, then find those lines on the page image. A deposition line that stays in order and matches the picture is sound. An affidavit line that arrives as scattered letters needs recognition again.

Can you see the text layer in a PDF without special software?

Yes. Any PDF reader and a plain text editor are enough, and there are three ways to do it: select the text on the page to see where the layer sits, paste a page into a text editor to read what it says, or save the whole file as plain text to read all of it.

The first method needs one caution. Highlighting shows that text exists. It does not show what the text says. A page can highlight neatly, line by line, over a layer that reads as nonsense. The post on why searchable does not mean correct explains that gap.

If nothing highlights and the cursor draws a box as though you were cropping a photo, the page has no layer to see. It needs OCR before any of this applies.

Copying a page into a text editor to read its text layer

Copying one page into a plain text editor is the quick method. It shows you the layer's actual characters, stripped of the picture that was making them look right.

1. Open the PDF and go to the page you care about.
2. Click once inside the page, then press Ctrl+A (Command+A on a Mac). Some readers select the whole document, which is fine for a short file. For a long one, drag from the top of the page to the bottom instead.
3. Press Ctrl+C to copy.
4. Open a plain text editor: Notepad on Windows, TextEdit in plain text mode on a Mac. Avoid a word processor, which may autocorrect what you paste and hide the errors you are looking for.
5. Paste. Set the editor to a fixed-width font so "l", "1" and "I" look different from one another.
6. Put the editor beside the PDF and read the two together.

What to read for

Start with what gets searched: names, amounts, dates, reference numbers. Then check three structural things. Do the lines come in the order a person reads them? Is anything on the page missing from the paste, such as a footer, a margin note or a stamp? Is anything in the paste that is not on the page, such as a second copy of the same paragraph, which points to two layers in one file?

If the copy command is unavailable, the file may carry a restriction set by its author. Ask whoever sent it for an unrestricted copy.

Exporting the whole text layer to a plain text file

Exporting the layer to a text file is the full-file method. PDF programs that can save a document as plain text write out the stored characters page after page, which gives you one file you can read, search and keep beside the PDF.

Look in your PDF program's Save As, Export or Convert menu for a plain text or .txt option. The name and place of the command differ between programs and versions, so check the program's own help if you cannot find it. Open the result in a text editor with word wrap turned on.

Then use it:

• Search it for the matter's names. Every spelling, including the wrong ones you suspect. A text editor's search shows you each occurrence in context.
• Skim for runs of rubbish. Strings of punctuation, single letters spaced out, or lines with no vowels mark the pages where recognition failed.
• Compare lengths. A 60-page file whose text export holds a few hundred words has pages with nothing behind them.
• Keep it. The export is a snapshot of the layer on the day you tested it.

An export that is not the layer

Be clear about which text you are getting. A PDF program's "save as text" writes out the layer already in the file. Uploading the PDF to an OCR service and downloading plain text is a different thing: what comes back is the service's own reading of your pages, and nothing promises that it matches the layer already in the file. Schema Law's plain text export, one of the eight formats listed on the homepage, is text Schema Law read from your upload. Such a reading is useful as a second opinion to hold against the existing layer. It tells you nothing about what the existing layer contains.

What does a healthy text layer look like next to a poor one?

A healthy layer reads like the page: same words, same order, with at worst a stray character. A poor one has wrong letters in names and numbers, broken or shuffled lines, or gaps. Set the two beside the picture and the difference is obvious within a few lines.

The visible page, from a lease amendment between fictional parties:

3. Base Rent. Commencing 1 March 2025, Tenant shall pay
Landlord $18,450.00 per month. Halvorsen Properties LLC
may adjust Base Rent under Section 4.2.

A healthy layer for that page:

3. Base Rent. Commencing 1 March 2025, Tenant shall pay
Landlord $18,450.00 per month. Halvorsen Properties LLC
may adjust Base Rent under Section 4.2.

A poor layer for the same page:

3. Base Rent. Cornmencing l March 2O25, Tenant shal1 pay
may adjust Base Rent under Section 4.Z.
Landlord S18,45O.OO per rnonth. Ha1vorsen Properties LLC

Read the poor one closely. "Cornmencing" and "rnonth" swap "m" for "rn". The year and the rent have capital O in place of zero, so a search for 18,450 fails. "Ha1vorsen" hides the landlord from a name search. Lines two and three have changed places, so a copied quotation would come out in the wrong order. On screen, both versions of this page look identical.

A third condition is worth recognizing: a layer that is partly healthy. Typed paragraphs come through clean while a handwritten date, a stamped "RECEIVED" or a table of figures is absent or garbled. Judge a layer by its worst important passage, not its average.

When the errors are few and they matter, the post on correcting OCR text in a PDF covers fixing a wrong word and keeping a record of the change.

Why does the text layer sometimes show through or print on the page?

The layer shows when something draws text that was meant to stay invisible. The hidden words are real text carrying an instruction not to paint them. A program that ignores the instruction, a conversion that drops it, or a tool that hid the text some other way can leave faint or doubled lettering on the page.

Causes to test for, one at a time:

• The viewing program. Open the file in a second PDF reader. If the ghost text appears in one and not the other, the file is fine and the first program is rendering it oddly.
• A conversion step. Printing to PDF, saving to another format and back, or passing the file through a portal can rewrite page content. Compare the file before and after each step you ran.
• Text hidden by color or size. A layer can be "hidden" by drawing it in white or at a tiny size instead of using the invisible mode. White text shows on a gray or colored scan.
• Text placed over the picture. If the layer sits on top of the image and is slightly visible, you see a shadow of each word, offset from the scanned letters.
• Selection left on. Highlighted text that was selected and never deselected can look like a tint over the page. Click an empty area.

If the ghosting appears on paper only, print one page from a different reader and with the "print as image" option if your printer dialog has one, then compare.

Visible layer text matters beyond tidiness. Words that nobody chose to put on the page are now on it, and they may be wrong words. For a filing or a production, go back to the last copy that displayed cleanly. The post on the OCR text layer as hidden text covers why that layer deserves a look before any PDF leaves the firm.