tubakhxn/sign_to_sentence — reverse-engineered prompt

Build me a simple real time webcam app called Tuba’s Glasses that can watch a person signing and turn a small set of hand signs into spoken English. I want it to recognize the main signs USE, TECHNOLOGY, FOR, HELP, NOT, and WAR, plus a few extra gestures if that fits, and show the live camera feed with hand tracking and face tracking on screen. When someone holds a sign steady for about 1.5 seconds, lock in that word and speak it out loud. If they raise both hands open for about 2 seconds, read the full interpreted sentence aloud.

I also want an emotion detection overlay on the video, just as a visual extra. Keep it practical and easy to run on a normal computer with a webcam, and make it install anything missing if possible. If anything is unclear, look up the current docs online and make sensible choices so it works end to end.