The end goal: Reverse-engineers the Wispr Flow desktop client and exposes its transcription and command APIs through a clean Python interface. Send audio files directly from Python, stream live audio, customize transcription behavior, and receive structured results — no UI interaction required.