musubipapi/computah ? reverse-engineered prompt
Reverse engineered prompt
Build me a Mac app that lets me control my computer with my voice.
It should listen for speech, turn it into text, figure out what I want to do, and then interact with the current app using macOS Accessibility controls. I want it to be able to open apps or websites, click buttons, type into fields, and adjust controls, then check whether the action actually worked before moving on.
Use Deepgram for speech to text and TypeSafe’s Jev to interpret the command and choose the next action. I also want a Debug Mode where I can type a command instead of speaking, see what the app is doing, and review the result after it runs. Make it careful about permissions, app state, and uncertain actions, and stop if it cannot confirm something instead of guessing. It should feel like an experimental but usable voice assistant for the Mac. Look up current docs online if you need to.
Are you gonna build this?
make sure you review the code using coderabbit