How to make a scanned PDF searchable on a Mac

A scanned PDF becomes searchable when OCR (optical character recognition) reads each page image and adds the words as an invisible text layer behind it. On a Mac, open the scan in Better PDF, choose Recognise, then File › Save a Digital Copy…. It runs on your Mac with Apple Vision, and nothing is uploaded.

What does “searchable” actually mean?

A PDF exported from a word processor or a web page holds text: characters, fonts and positions. You can search it, select it and copy it.

A PDF from a scanner or a phone camera holds pictures. Each page is a photograph of a sheet of paper. It looks like text, but to the computer it is pixels, which is why ⌘F finds nothing and why dragging across a line selects the whole page instead of a word.

OCR looks at the picture, works out which shapes are letters, and writes down the words with their positions on the page. A searchable PDF keeps the original image on top and puts those words underneath, in a layer drawn with no visible ink. You still see the scan exactly as it was. Search, selection, copy and paste, screen readers and Spotlight see the hidden words.

Because the layer is lined up with the image, a search highlight lands on the word in the scan. When the alignment is off, which happens with crooked pages, the highlight lands a little beside it.

Why do language and scan quality matter?

OCR is a guess with good odds. Two things shift the odds more than anything else.

Language. A recogniser that knows it is reading German expects “ß” and “ü” and words like “Mietvertrag”. Told it is reading English, it will turn them into something close and wrong. Better PDF reads in the languages listed in System Settings › General › Language & Region, in their order, so if you scan documents in a second language, add it there. Apple Vision covers English, the major European languages and several Asian ones; the exact list depends on your version of macOS.

Page quality. Most errors come from the scan rather than the recogniser:

• Resolution. 300 dpi is the usual target. Below about 200 dpi, small print and punctuation start to go. Better PDF renders each page at about 300 dpi before reading it, but it cannot add detail the scan never had.
• Skew and curl. A page photographed at an angle, or a book page curving into the spine, gives bent lines that read badly.
• Contrast. Faint photocopies, grey-on-grey letterheads, shadows from a phone held over the page.
• Handwriting. Neat block capitals often work. Joined-up handwriting mostly does not.

If a scan reads badly, scanning again is usually faster than correcting the result.

How to make a scan searchable in Better PDF

You need macOS 26 or later.

1. Open the scanned PDF with ⌘O. A page with no text and an image covering it counts as a scan, and Better PDF shows a card offering Recognise or Not now. You can also choose File › Recognise Text at any time.
2. Choose Recognise. Only the scanned pages are processed; pages that already have text are left alone. The text is usually ready in seconds.
3. Search with ⌘F. Every match is listed by page, and you can select and copy text straight off the scan.
4. To keep the text, choose File › Save a Digital Copy…. Better PDF writes a new file named <name> (digital).pdf with the original page images and an invisible text layer under them. Metadata, bookmarks and annotations come along.

The original file is never modified. Until you save a digital copy, the recognised text belongs to the open document in that session only, which is handy when you just need to find one clause and close it again.

If a page takes too long, Better PDF retries it with a faster recogniser instead of making you wait. The fast pass is a little less accurate, so it is worth a glance at the pages that got it.

On the Mac, Recognise and Save a Digital Copy are part of Pro, and the seven-day trial on first launch includes them. The short version is in Help: Recognise a scanned PDF.

What can the built-in apps do?

Some of it, and it depends on what you need.

Recent versions of macOS can recognise text in images, so in the built-in PDF viewer you can often select and copy words from a scanned page. That is real OCR, done on your Mac, and for copying a single address off a scan it is enough.

What it does not do is save that text into the file. Close the document and the recognition is gone. Spotlight cannot find the PDF by its contents, a colleague opening it in another app sees a picture, and searching a hundred pages is hit and miss.

So: the built-in viewer for a quick copy. A digital copy with a text layer when the file needs to be searchable wherever it goes.

Does any of this leave your Mac?

No. Apple Vision runs on the device, and Better PDF does not send pages anywhere to be read. Recognition works with Wi-Fi turned off. That matters for the documents people most often scan, which tend to be the private ones: contracts, letters from a doctor, passports, tax papers. We wrote about the general question in Is it safe to upload a PDF to an online converter?

The Convert tools can also recognise scanned pages on the way out, so a scan can go straight to text, Markdown or Word without a separate step. And if the paper is still on your desk, Scan mode skips the scanner: point an iPhone at the page through Continuity Camera, or drop in photos, and Better PDF finds the edges, squares the perspective, cleans it up with a filter, recognises the text and stitches the pages into one searchable PDF. On iPhone and iPad, Lens does the same on the phone, with recognition included for free.

Better PDF needs macOS 26 or later. Download it here.

Questions people ask

Why can't I search my scanned PDF?

Because each page is an image of text, not text. Search needs characters, and a scan has none until an OCR step adds them.

Does OCR change how the PDF looks?

It should not. A searchable PDF keeps the original page images and puts the recognised words in an invisible layer underneath, lined up with the words you see.

Can I copy text from a scan without making it searchable?

Often yes, because recent versions of macOS can pick out text in images and let you copy it. That text is not saved into the PDF, though, so Spotlight and other apps still see a picture.

How accurate is on-device OCR?

On a clean, straight scan of printed text at around 300 dpi, very accurate. Handwriting, faint photocopies, small print and mixed languages are where errors creep in, so check names and numbers before relying on them.