This is probably going to be a WONTFIX. In order for Google OCR to work well, it needs to know what language it's reading. (You would be surprised at how bad it is at guessing.) Currently when you click the Google OCR button, it passes the language code for that wiki (under the assumption that most text on that wiki is going to be in the content language. We could add another step where it asks for the language explicitly, but that would needlessly slow down the process for the 99% of cases where the language of the text matches the language of the wiki. One workaround is to use https://tools.wmflabs.org/ws-google-ocr/ in those cases, since it allows you to specify the language of the text.

I'll leave the ultimate decision up to @Samwilson, but I would recommend WONTFIX.

We could add an extra button to the toolbar that just opens the ws-google-ocr tool in a new tab, with the image URL pre-filled. That would probably be the quickest to implement. For the WikiEditor toolbar, it could go a level down, under 'proofreading tools'; for the old toolbar, it'd just be next to the existing button (with some sort of external-link icon?).

I don't really have an idea of how many works there that contain lengthy sections of other languages, where this feature would be useful. Might ask on the mailing list.

(I'm happy to work on this in my own time; we've enough CommTech stuff going on I think.)

@Samwilson Is this the ticket you mentioned in the standup?

No, this one: T201282: Google OCR button gadget on English Wikisource sometimes does not work (and looks weird)

I think this issue is resolved now with the new OCR button. It's now possible to open the 'advanced options' and enter different (or multiple) language names.

@Bodhisattwa can you confirm?

In T198040#7223168, @Samwilson wrote:

I think this issue is resolved now with the new OCR button. It's now possible to open the 'advanced options' and enter different (or multiple) language names.

@Bodhisattwa can you confirm?

Yes, its done now.