WikimediaOCR

A tool to transcribe text from scanned images on Wikimedia Commons, for use on Wikisource and elsewhere.

Transcribe an image

Insert an image URL hosted on a Wikimedia server such as: upload.wikimedia.org, upload.wikimedia.beta.wmflabs.org and thumb.wikimedia.org

Text printed on the reverse of a thin page can show through and be read as characters that aren't there. These options remove it before the image is sent to the engine. They help most on pages with large blank areas, and tend to make densely printed pages slightly worse.

Reduce show-through
Clears the background but leaves the ink at its original shade, so strokes keep their weight.
Black and white
Reduces the page to pure black and white. Removes the most show-through, at some cost to fine strokes.