I was using it by 2020 for sure, so it predates the macOS and iOS feature. This was most handy in ERP software we were using that had most info display in unselectable windows. Really annoying when you wanted to copy something like a part number or invoice and put it in an email. This got us around that, and when macOS added the feature it still didn’t help us since these weren’t images.
The Power Toys link says it’s based on Joe Finney’s Text Grab, and at the bottom of its GitHub page it links to the TextSniper app as the Mac version, with an affiliate link. I’m guessing that means the Mac app was inspired by the Windows program.
No, this predates having it on either iOS or macOS by a year or two. I still found it more useful because this doesn’t require using images; the vast majority of my usage was when working for a company that had stupid ERP software where much of the data was displayed onscreen but couldn’t be copied.
I think on macOS and iOS it only works in actual image files, but this tool predates that by a year or two. This does the same thing but doesn’t require an image file; you just press the shortcut on your keyboard, draw a box over whatever’s on your screen that you want, and the text in the box goes on your clipboard. I think it’s effectively taking a screenshot but not saving it to disk, so you don’t have to clean those up later.
I have a handy little app on macOS called TextSniper that takes a screenshot of a selected area, then runs OCR on that screenshot and puts the text on the clipboard. It’s perhaps the most useful $10 I’ve ever spent and I’m frankly surprised this doesn’t exist on other systems. A year or two after this was released Apple started letting people copy text directly out of images, so they might do the usual Apple thing of killing it by directly adding it to the OS. There might be something like this on Linux by now but I haven’t heard of it on Windows.
FYI this article is from June 2024
They had a fairly detailed blog post a few days or weeks ago about the rollout. The plan is to bring it to open source but they’re still working on issues with it that are easier to control on their own servers. IIRC the code for it is actually in the open source version but disabled. I think they said if you know what you’re doing you could go into the code to enable it but it’s unusably slow right now, or something like that.