That will do the trick, even though its a bit hackish... I don't know if people would like this but it might be useful to include a snippet: {Image stripped, 123 bytes} or some other information to the text file explaining that an image used to be here?

I agree that it's a specific and not-at-all pretty fix. I'm just not familiar enough with pyth and the finer points of the RTF format to intelligently make changes to the design.

As for the snippet, I do a lot of content analysis, and I use pyth to process RTFs into plain text. It's probably my specific research use case, but I'm wary of adding text into a document. Also, the images in my documents are an artifact of the data provider (not the original data). It may be a good option, though. If I were looking at documents with "real" embedded images, being able to capture that fact might lead to interesting results. I would guess that a lot of use cases would similarly be interested in at least knowing about images.