Orihako PDF

Text recognition (OCR)

Read the text of a scanned PDF so you can search and copy it. The text is also saved as a text file.

Recognition is not 100% accurate: handwriting, faint or skewed scans and very small text cause mistakes. Check the text before you rely on search or copy.

Your files are never uploaded.

How to use it

  1. Drop the scanned PDF.
  2. Choose the language(s) of the document and, if you want, the pages.
  3. Press Run. Each page is read on this device; then download the searchable PDF and the text file.

Read the guide: Make a scanned PDF searchable, without uploading it

Questions and answers

Is the scan sent to a server for recognition?

No. The text recognition runs inside this page. The reading data is downloaded from this site the first time and kept on your device.

How accurate is it?

Good on clean, straight scans of printed text. On our own Japanese test pages (invoices, estimates, contracts, two-column and vertical pages), about 1 to 2 characters in 100 were wrong in horizontal text and about 4 in 100 in vertical text. Tables drawn with lines are read cell by cell. Weaker on handwriting and on faded, fax-like scans.

Why is the first run slow?

The first time, the reading data has to be downloaded (about 18 MB for Japanese, a few MB for English or French): the page shows its size and asks first. After that a page takes a few seconds to a little over ten, depending on the computer or device.

Recipes

Other tools

All tools