ScreenTutor

My mom teaches music at a middle school in Vietnam. Every week her principal sends her tasks on Zalo: download this file, fill in that form, export the lesson plan to PDF, print it. Each one is a chain of clicks she doesn't know. I live in Buffalo, so I can't sit next to her. ScreenTutor sits next to her instead. She holds Ctrl+Alt (or presses a round button) and asks in Vietnamese. It looks at her screen and says the next step out loud, and a dot flies to the exact thing to click as the voice starts. She does the click herself. It never clicks or types for her. After each click it checks the new screen to see whether the step and the task are done, then goes on. - A big task is split into parts, shown on a small "dynamic island" card. - A step that sends or replaces something turns amber and says what will happen. - If she clicks ahead while it is thinking, it sees that the screen has moved on and says nothing about the old one. - If she says "stop", it stops. It reads pixels, not app APIs, so it works with any Windows app: Zalo, Word, Brave, and MuseScore, etc..., which it had never seen and has no code for.