Which languages are supported?

Over one hundred, including English, Spanish, French, German, Arabic, Hindi, Japanese, Korean and Chinese. Pick the language actually printed on the page for best accuracy.

On a clean 300 DPI scan, typically 97–99% character accuracy. Handwriting and low-contrast photos fare much worse.

Why is the file bigger afterwards?

The invisible text layer adds a font subset and positioning data. The scanned image is unchanged.

ream.

runs offlineOpen a tool

PDF · free · no account

OCR a scanned PDF

A scan is a picture of words. Optical character recognition turns it into a document you can select, search and copy — with the original image untouched underneath and an invisible text layer on top.

100+ languages
Hidden text layer
Original image kept
Runs locally

Job ticket
In: PDF
Out: One file
Runs: In this tab
Upload: None
Account: Not required

1 · Choose a file

Drop your file here

or browse your device

PDF filesOne filenever uploaded

or paste a file with ⌘V

2 · Settings

Document language

Pick the language printed on the page, not your interface language.

Recognition DPI

DPI

Add a file first.

Live network monitor

watching

Files sent: 0
Uploaded: 0 B
Foreign hosts: 0

measured by your browser’s resource timing API

OCR PDF: questions worth answering

Which languages are supported?: Over one hundred, including English, Spanish, French, German, Arabic, Hindi, Japanese, Korean and Chinese. Pick the language actually printed on the page for best accuracy.
How accurate is it?: On a clean 300 DPI scan, typically 97–99% character accuracy. Handwriting and low-contrast photos fare much worse.
Why is the file bigger afterwards?: The invisible text layer adds a font subset and positioning data. The scanned image is unchanged.

Keep going

all 34 tools