sikandarmoyaldev-rvc

Lightweight, WebUI-free RVC (Retrieval-based Voice Conversion) inference engine for the python project.

📦 Installation

pip install sikandarmoyaldev-rvc
# or
pip install git+https://github.com/sikandarmoyaldev/python-rvc.git

📖 Usage

1. Download Base Models

Before converting, you must download the required base models (HuBERT, RMVPE, etc.).

from rvc import download_base_models

download_base_models("assets")

2. Convert Audio

Initialize the converter with your custom .pth model and convert your audio files.

from rvc import RVCConverter

# Initialize with your model path and assets directory
converter = RVCConverter(model_path="alex.pth", assets_dir="assets")

# Convert a single file (output defaults to "output/input_filename.wav")
output_file = converter.convert(
    input_path="group-1.wav", pitch=2, f0_method="rmvpe", index_rate=0.0
)

print(f"Saved to: {output_file}")

🚀 Development

This project uses uv for dependency management.

# Install dependencies
uv sync

# Format and lint code
uv run ruff format .
uv run ruff check . --fix