NeoRVC 🎵

---

---

NeoRVC is an enhanced extension of AIcoverGen, designed to create AI-generated song covers using any RVC v2 trained voice from YouTube videos or local audio files. Perfect for developers integrating singing functionality into AI assistants, chatbots, or VTubers, and for fans who want their favorite characters to sing their favorite songs.

This project will rarely get updates because I'm tired of working alone, open PR if you want improve this project.

Features

• User-Friendly WebUI: Simplified voice model conversions and downloads.
• Advanced Pitch Detection: Supports hybrid methods (rmvpe, crepe, fcpe) for precise vocal conversions.
• Flexible Input Options: Generate covers from YouTube URLs or local audio files.
• Customizable Outputs: Choose between full song or vocals-only output, with options for wav or mp3 formats.
• Voice Model Management: Download public voice models with search/tag filters or upload locally trained RVC v2 models.
• Enhanced Audio Controls: Adjust pitch, reverb, volume for main/backup vocals and instrumentals, and apply advanced RVC settings (filter radius, RMS mix rate, protect).
• CLI Support: Perform inference and model downloads via command-line interface.
• Local Network Sharing: Share WebUI over a local network for collaborative use.
• Intermediate File Retention: Option to keep isolated vocals/instrumentals for further editing.

Changelog

• Added WebUI for streamlined conversions and voice model downloads.
• Implemented hybrid pitch detection with rmvpe, crepe, and fcpe methods.
• Enabled support for local audio file inputs.
• Introduced option to retain intermediate files (e.g., isolated vocals/instrumentals).
• Added support for Pixeldrain download links for voice models.
• Improved vocal conversions with new rmvpe pitch extraction technique.
• Included volume controls for main vocals, backup vocals, and instrumentals.
• Added index rate, reverb, and advanced RVC options (filter_radius, rms_mix_rate, protect).
• Enabled local file uploads via file browser in WebUI.
• Supported pitch change and autotune for vocals.
• Introduced CLI inference for automated workflows.

CLI Usage

NeoRVC provides a powerful command-line interface for inference and model downloading. below are the available commands and their options:

Command	Description	Usage Example
infer	Generate AI song cover using a YouTube URL or local audio file.	python ./neorvc/main_cli.py infer "https://youtube.com/watch?v=example" "my_voice_model" --pitch_change 2 --keep_files --output_type full
download	Download an RVC voice model from a specified URL.	python ./neorvc/main_cli.py download "model_url" "my_voice_model"

CLI Options for infer

Option	Description	Example Value
--pitch_change	Pitch change in semitones	2
--keep_files	Keep intermediate files	N/A
--output_type	Output type: full or vocals	full
--main_gain	Main vocals gain (dB)	0
--backup_gain	Backup vocals gain (dB)	0
--inst_gain	Instrumental gain (dB)	0
--index_rate	Index rate for voice conversion	0.5
--filter_radius	Filter radius for voice conversion	3
--rms_mix_rate	RMS mix rate	0.25
--f0_method	F0 extraction method (e.g., rmvpe, crepe, fcpe)	rmvpe
--crepe_hop_length	CREPE hop length	160
--protect	Protect voiceless consonants	0.33
--output_format	Output format (e.g., mp3, wav)	wav
--f0_autotune	Enable F0 autotune	N/A
--f0_autotune_strength	F0 autotune strength	0.5

CLI Options for download

Option	Description	Example Value
model_url	URL for RVC model	https://example.com/model.zip
voice_model	Name of the RVC voice model	my_voice_model

References

NeoRVC builds upon the following projects:

• Vocoders:

  • HiFi-GAN by jik876
  • Vocos by gemelo-ai
  • BigVGAN by NVIDIA
  • BigVSAN by sony
  • vocoders by reppy4620
  • vocoder by fishaudio

• VC Clients:

  • Retrieval-based-Voice-Conversion-WebUI by RVC-Project
  • rvc-cli by blaisewf
  • So-Vits-SVC by svc-develop-team
  • Mangio-RVC-Fork by Mangio621
  • VITS by jaywalnut310
  • Harmonify by Eempostor
  • rvc-trainer by thepowerfuldeez
  • AICoverGen by SociallyIneptWeeb

• Pitch Extractors:

  • RMVPE by Dream-High
  • torchfcpe by CNChTu
  • torchcrepe by maxrmorrison
  • anyf0 by SoulMelody

• Other:

  • FAIRSEQ by facebookresearch
  • FAISS by facebookresearch
  • ContentVec by auspicious3000
  • audio-slicer by openvpi
  • python-audio-separator by nomadkaraoke

We extend our gratitude to the authors and communities behind these projects for their invaluable contributions.

Contributing

We welcome contributions to NeoRVC! To get involved:

• Report issues at GitHub Issues.
• Submit improvements via Pull Requests.
• Star the repository to show your support! 💙.

Support the project via donations: