OCR a PDF on a Mac
without paying for anything.

Aug 13, 2026 · 3 min read

macOS does this on its own. Open the PDF in Preview, choose File then Export, and tick Embed Text. Preview recognises the text and writes it into the file, so search works afterwards in any reader. The option is there from macOS 14 Sonoma, and it only appears when the PDF has no text in it yet.

Preview, Export, Embed Text

Open the scanned PDF in Preview. Choose File then Export, and tick Embed Text in the panel that opens. Preview reads every page and writes the recognised text into the file it saves.

After that the document behaves like any normal PDF. Command F finds words, text selects properly, and it stays that way for anyone you send it to.

Two things to know. The option arrived in macOS 14 Sonoma, so an older Mac will not show it. And it only appears when the document has no recognised text yet, which is why it goes missing on files that are already fine.

Live Text is not the same thing

Since macOS 13 Ventura, hovering over a scanned page in Preview turns the cursor into a text cursor and lets you drag across the words and copy them. That is Live Text, and it fools a lot of people into thinking the file was converted.

It was not. Live Text happens in the viewer and writes nothing back. Close the document, search it, or send it to somebody else, and there is still no text in there.

That is what almost every "Preview OCR not working" post turns out to be. Nothing was broken. Nothing was ever written.

Live Text also works on the page in front of you rather than on the whole file, so it will not make a two hundred page report searchable no matter how long you scroll.

Live TextEmbed Text (Export)
SincemacOS 13 VenturamacOS 14 Sonoma
What it doesLets you select and copy text on the page you're viewingWrites the recognised text into the saved file
CoversOnly the page in front of youEvery page in the document
Survives closing the file?NoYes
⌘F search works?NoYes

Three cases Preview will not cover

A folder of files. Preview does one document at a time, so batches want a dedicated tool.

Handwriting. Recognition on handwriting is unreliable everywhere, and macOS is no exception.

And the case nobody lists: you did not want to convert anything. You wanted to know what one line said.

Reading a line without converting the file

Converting a document to get at one sentence is a lot of ceremony, and it is no help whatever if the thing you are reading is a locked file, a preview thumbnail, or a page in an app that is not Preview.

Leafy takes the other route. Press ⌥A and drag a box around the line. macOS reads that patch of screen and gives you the translation next to the original, and any word in it can be looked up in the sense the sentence is using and saved with that sentence attached.

Nothing gets converted, nothing gets uploaded, and your reader stays exactly where it was. It is the wrong tool for turning a book into a text file and the right one for getting through a paper in a language you are still learning.

Where the text goes

The page image is read on your Mac and never leaves it. The recognised line does go out to an AI to be translated or explained, with nothing attached that says who you are.

Your library stays on your machine. No account, no sync server, no ads. The privacy page lists exactly what is sent and what is kept.