AUv3 RVC Voice Converter

I need a production-ready AUv3 that puts Retrieval-based Voice Conversion (RVC) directly inside Logic Pro and any other AUv3 host on both macOS and iOS. Real-time voice conversion is essential, so latency must stay low enough for live tracking and monitoring at 48 kHz. Alongside that, I want the same engine to handle batch processing of audio files—for example, dragging a folder of takes onto the plug-in window and letting it render them offline.

Voice models should be fully customizable: the plug-in must ship with a small library of pre-defined voices yet also allow users to load—or even train—their own RVC models without leaving the session. A simple browser for selecting, tagging, and updating models would be ideal.

Deliverables
• Signed, notarised AUv3 binary that runs natively on Apple Silicon and Intel Macs, plus a universal build for iOS/iPadOS hosts such as GarageBand.
• Xcode project with clear build instructions so I can compile future versions myself.
• UI/UX that exposes real-time / batch modes, model management, and basic mix controls (wet/dry, formant shift, gain).
• Documentation covering installation, parameter automation in Logic Pro, and the steps required for users to import or train additional models.

Once the plug-in loads in Logic Pro and performs reliable, low-latency conversion using both the included and user-supplied voices, the job is complete.