World-renowned open-source retrieval-based voice conversion framework widely used for AI song covers and live streaming.
Fully open-source on GitHub with easy WebUI installers; unlimited local model training, song covers, and live voice change
RVC (Retrieval-based Voice Conversion) is the preeminent open-source neural voice transformation toolkit powering viral AI music covers and real-time streaming worldwide. Built on top of customized VITS architectures with Faiss vector retrieval, RVC replaces source vocal timbres while strictly preserving original pitch curves, vocal vibratos, and emotional dynamics. Training a high-fidelity personalized voice model requires only minutes of clean audio on consumer hardware, making it the industry benchmark for vocal transformations.