GUIDE · 22 NOVEMBER 2026

How to OCR Scanned Documents into Searchable, Copyable Text

Learn how to OCR scanned documents on a Mac: pick the right method, get cleaner scans, and turn PDFs into searchable or editable text.

Last updated: 22 November 2026

To OCR scanned documents, run them through optical character recognition (OCR) software. It reads the pixels of each page and adds real, selectable text, either as a hidden layer behind the image (a searchable PDF) or as a separate text or Word file. On a Mac you can do a quick copy with Live Text, or use a dedicated OCR tool when you need a whole searchable PDF or an editable document.

What does OCR actually do to a scanned document?

OCR turns a picture of text into text a computer understands. A scanned PDF is only a stack of photos, so you can't search it, select a sentence or paste a paragraph. OCR recognises the letters in each image and stores them as text.

What you get depends on the output you choose:

  • Searchable PDF: the page still looks exactly like the scan. An invisible text layer sits over it, so search, highlighting and copying work.
  • Plain text: just the words, with no layout. This suits pasting into notes or an email.
  • Editable document (such as Word): the text is reflowed so you can edit it, but layout may shift.

OCR does not understand meaning, and it makes mistakes. Treat its output as a very good first draft, not a certified copy. That matters for contracts, invoices and anything with numbers.

Which OCR method should you use on a Mac?

Use Live Text for a one-off copy from a photo, a searchable-PDF OCR pass for archives you need to search, and a conversion to Word or plain text when you plan to edit. The right choice depends on what you need afterwards, not on the file type.

Your goalBest methodOutputWatch out for
Copy a paragraph from a photo or screenshotLive Text in Preview or PhotosText on your clipboardOne image at a time
Search a 40-page scanned PDFOCR to searchable PDFSame-looking PDF with a text layerWrong language setting gives gibberish
Edit a scanned letter or reportOCR to Word or plain textEditable fileTables and columns may need tidying
Process a folder of scansBatch OCR toolMany searchable filesCheck a sample before trusting all
Sensitive documents (medical, legal, financial)Local OCR, not an upload siteAny of the aboveOnline services require sending the file

Live Text: the free quick option

macOS has OCR built in. According to Apple's Preview guide, you can drag to select text, Control-click the selection, then choose Copy Text in a photo. It works for short snippets, but it won't give you a searchable version of a 60-page PDF.

Dedicated OCR: for whole documents

When the goal is to find words across an entire scan or turn it into a file you can edit, a tool that processes whole PDFs and batches is more practical. Many popular guides point to online services for this, which means uploading the document. We cover that trade-off in our comparison of offline vs online file converters.

How do you get the best OCR results from a scan?

Good OCR starts at the scanner, not the software. Scan at 300 DPI or higher, keep the page flat and straight, and use even light. Poor input is the main reason for errors, and no software fully recovers detail that was never captured.

Resolution is the most common culprit. One accuracy guide from LlamaIndex says anything below 300 DPI causes a measurable drop in character recognition accuracy, and a research paper on OCR preprocessing notes that current OCR software is designed for recognising printed text at 300 dpi.

A pre-scan checklist:

  • Set the scanner to 300 DPI for normal print. Use 400–600 for very small type, such as footnotes or receipts.
  • Use black and white or greyscale for plain text pages. Colour is only needed if you want to keep it visually.
  • Flatten the page. Curved book spines and folds distort letters near the edge.
  • Straighten it. Crooked scans reduce accuracy, and deskewing helps.
  • Avoid shadows when using a phone. Shoot from directly above, in daylight or under even light, and check for blur before moving on.
  • Crop to the page. Dark borders and desk backgrounds can confuse the layout detection.
  • Scan one document per file. It's easier to name, split and verify later. If you've scanned many things into one PDF, you can split the PDF and extract specific pages first.

What if all you have is a low-quality photo?

Enlarging a small image before OCR can help a little, but it can't create sharpness. The same research paper found that images down to about 150 dpi could be enlarged to 300 dpi and still give excellent results on printed text. Below roughly 100 DPI, expect significant errors. If you can rescan, rescan.

How do you make a scanned PDF searchable?

Open the PDF in an OCR tool, choose the document's language, select a searchable-PDF output, and run it. The result looks the same but has a text layer you can search with ⌘F.

The steps are the same in most tools:

  • Add the file. Drop in the scanned PDF or image.
  • Pick the recognition language. This is the setting people most often miss. Recognising a German letter as English produces broken umlauts and odd words.
  • Choose the output: searchable PDF, plain text, or Word.
  • Choose where to save, ideally not over your original.
  • Run it, then open the result and search for a word you know appears on the page.

Convert2Any has this built in. Its site lists text recognition for scanned PDFs and pictures in the languages you choose is not quoted from a source, so to be exact: the app's own page says it recognises text in scanned PDFs and pictures in the languages you choose, makes PDFs searchable, and lists outputs including searchable PDFs, image to text, image to Word and a plain-text copy, with 100+ languages. Processing happens on your computer.

Verify the result in two minutes

Don't assume the text layer is correct. Check it:

  • Search for three words you can see on different pages, including one near the bottom of a page.
  • Select a paragraph and paste it into a text editor to see whether lines and characters came through.
  • Check numbers closely. Confusions such as 0 and O, 1 and l, or 5 and S are classic OCR errors.
  • Look at names and dates in anything legally or financially important and compare them with the original.

How do you turn scanned text into something you can edit?

Choose Word or plain text as the output instead of a searchable PDF. Expect to spend a few minutes fixing line breaks, columns and tables, because OCR recovers words much more reliably than complex layouts.

Simple single-column pages, such as letters, contracts and book pages, convert well. Multi-column newsletters, forms and tables are harder, because the software has to decide reading order. Photos of receipts with crumpled paper are harder still.

Practical tips:

  • For layout-heavy documents, keep the searchable PDF as your master and extract only the text you need.
  • For tables, copy the numbers and rebuild the table, or check each cell against the scan.
  • Keep the original scan. If a mistake turns up later, you can go back to the source.
  • Handwriting is a different problem. Standard OCR is designed for printed text, and handwriting recognition is far less reliable. Don't count on it for important notes.

If your scans are photos in formats like HEIC, you may also want to convert them first; see how to convert HEIC to JPG on a Mac in batches.

Is it safe to OCR sensitive documents online?

Online OCR means uploading your file to someone else's server. For passports, medical letters, contracts or tax documents, a local tool avoids that step. Whether a given service keeps or deletes uploads depends on its own policy, which you should read before using it.

Several popular OCR guides recommend upload-based tools. Adobe's online route, for example, has you to upload your scanned PDF and sign in to download the result. That's convenient, but it's a different privacy position from processing on your own machine.

A few related habits help:

  • OCR first, then redact. If you must hide names or numbers, do it properly afterward; see our guide to redacting a PDF so hidden text is really removed. Remember that OCR creates a new text layer, so redactions must cover it too.
  • Compress after OCR if the file is large. Scans at 300 DPI can get big, and our guide to compressing a PDF without losing readability explains how to shrink it without harming the text.
  • Name the files clearly so you can find them by search later.

Convert2Any runs conversions and recognition on your Mac and can handle batches of files, so it suits people who want searchable PDFs without sending documents elsewhere. It has a 3-day free trial with no account needed.

Frequently asked questions

Can I OCR a scanned PDF for free on a Mac?

Yes, for small jobs. Live Text in Preview lets you select and copy text from images and many scans. For whole documents or batches, you'll usually need a dedicated tool, and free online ones involve uploading your files.

Does OCR change how my PDF looks?

Not when you create a searchable PDF. The page image stays the same and the recognised text is added invisibly behind it. Converting to Word or plain text does change the look, because the text is reflowed.

Why does my OCR output contain strange characters?

The usual causes are the wrong recognition language, a low-resolution scan, skewed pages or a dark background. Rescan at 300 DPI or higher, straighten the page, and make sure the language matches the document.

Can OCR read handwriting?

Not reliably. Standard OCR is built for printed text, and handwriting varies too much for consistent results. Test on a sample and proofread everything if you try it.

How accurate is OCR?

It depends heavily on the scan. Clean, 300 DPI printed pages usually come out very well, while faded, skewed or low-resolution pages produce more errors. No universal accuracy figure is established, so always spot-check numbers and names.

Explore supported formats Back to Blog