w-okada/voice-changer — reverse-engineered prompt

Build me a real time AI voice changer app called VCClient. I want to speak into my microphone and hear my voice converted live with as little delay as possible, so it can be used for streaming, voice chat, games, or recording.

It should let me choose my microphone and speaker, load supported voice models like Beatrice and RVC when available, adjust things like pitch, formant, chunk size, and buffer settings, and show a simple visual indicator so I can tell audio is flowing. Please include shortcut keys for common actions and make it work either fully on my computer or by connecting to another machine on the network to do the heavy processing.

Make it usable on Windows, Mac M1, and Linux where possible. Also expose a simple REST API so other apps can control it. Include clear setup instructions, model license notes, and basic troubleshooting for audio device and connection problems. Look up current docs online if you need to.