Text recognition (OCR)

A scanned page is a picture until something reads it. Scanpaper’s text recognition turns the printed words on the page into text you can use, without sending the scan anywhere.

What you can do with it

Copy text out of a scan. Open a document, tap Text, and the recognized text for the page appears. Copy all of it, or share it as a TXT file. That’s the quick route for pulling an address off a letter or a paragraph out of a handout.

Export the whole document as text. Export as Text gives you a plain TXT file of the document, ready to paste into notes or an email.

Search your library by what’s inside. The search box in Documents looks at document names and at the text inside your scans. Search “invoice 2026” or a company name and the matching scans come up even if you never renamed them.

Send searchable PDFs. Once OCR has run, the PDFs you export carry an invisible text layer. The page still looks like your scan, but the recipient can search it, select a sentence and copy it.

How it works

Recognition happens on the phone, using the text engine built into the system: Apple Vision on iPhone, Google ML Kit on Android. There’s no upload step and no wait for a server, and it works in airplane mode.

It’s tuned for printed text: letters, forms, receipts, book pages, printed handouts. Handwriting isn’t something to rely on.

Getting cleaner text

Recognition is only as good as the scan. A few habits help:

• Let the edge detection crop the page, so there’s no table or carpet in the shot.
• Use the B&W or Auto filter for plain text pages; strong contrast helps the engine.
• For a thick book, use good light and press the spine flat, since curved lines read worse.

On the web

If you just need text from one image right now, the free image to text tool on this site runs in your browser. It’s a separate utility with its own engine, and it doesn’t connect to the app.

Text recognition is part of Scanpaper Pro, which has a free trial.