Really excited to try this one out.

Thank you for the post. I will definitely be testing it.

I tried it. It's got a long way to go before prime time. Not sure if it's due to limits set by #Gemini or what. But, Yes, you can tab through the elements, but not much else. Plus, it seems to take over windows when that screen is in view, to where I can't alt tab or anything until I close this app out. This after generating what's on the screen. Can't even use jaws navigation keys.

@serichsen sounds very interesting, can it be made to use local models? using a compatable open ai like api?