A screenshot of Glyphtracer


Glyphtracer takes an image that contains pictures of several letters. It recognizes all them and lets the user tag each letter to a Unicode code point. It then converts the images to vector form and writes them to a FontForge’s data format. The font can then be finalized with FontForge.


Glyphtracer requires PyQt5 and Potrace, which is an image vectorizer.

It has been only tested on Linux. It might work on OSX or Windows. It might not.

Usage tips

Glyphtracer only processes 1 bit images, but they can be in any format understood by Qt.

The letter recognition is based on white space. Thus every row must be separated from other rows by a continuous horizontal strip of white. If this is not the case (because, for example, your image is tilted), detection will fail. Similarly letters on a single row have to be separated by vertical white space. Just give your letters lots of “room” on all sides and everything will work.