AbhishekBarali/SpeakoFlow ? reverse-engineered prompt

Build me a desktop voice assistant called SpeakoFlow that works on Windows, macOS, and Linux. I want to press a hotkey and talk, then have my words typed into whatever app I am using, either live as I speak or all at once when I stop. It should also have a voice driven assistant panel where I can ask questions, get answers read back aloud, and optionally use screen context, web search, and personal memory if I turn those on.

Make it local first, so speech to text runs on my own machine and my voice does not leave my device unless I choose an external AI provider. Include a simple setup flow for choosing a transcription model and an assistant model. Add settings for changing hotkeys, trigger phrase, voice, tone, and assistant behavior.

Use the current project docs and build setup if you need to, and make it feel polished, privacy focused, and easy to use.