I know there are most certainly more polished solutions out there now, but as we wait for the ChatGPT desktop app with builtin vision support I figured I'd release this now under an MIT license for whoever might find it useful. Plus this just automates your browser, meaning there's no need to put in API tokens or the like (you do have to be logged into ChatGPT, however).
I don't think I'll put it on the Chrome Web Store, so the installation steps are a bit annoying, but if even one person finds use in this, I'll be happy!