litagin02/rvc-tts-webui

Combines edge-tts for baseline speech synthesis with RVC voice conversion models to enable custom voice cloning on CPU or GPU. The pipeline chains Microsoft's neural TTS engine output through retrieval-based voice conversion, allowing users to apply pre-trained RVC models (`.pth` + `.index` files) to synthesized speech. Built as a Gradio interface that supports batch model management and runs efficiently on modest hardware without GPU acceleration.

336 stars. No commits in the last 6 months.