iTzFrancesco/Traflix-Voice ? reverse-engineered prompt

Build me a Windows desktop voice dictation app that feels lightweight and private.

I want to hold or tap a global hotkey, speak into the mic, and have the app turn my speech into text and paste it into whatever app I’m currently using. By default it should work locally with Whisper so audio stays on my machine, but I also want an optional cloud mode with Groq if I choose to enable it in settings.

Please include a clean settings window, a small always on top overlay that shows recording status and a waveform while I’m speaking, system tray support, and a simple way to pick my microphone, language, model, and whether I want automatic paste on or off. It should support on demand model downloads, local history and usage stats, and basic CPU plus optional CUDA selection. If you need to look up current docs online for Tauri or Groq, go ahead.