Skip to content

pentarab/ComfyUI-For-ChatterBox

Repository files navigation

ComfyUI-For-ChatterBox

ComfyUI custom nodes for Chatterbox TTS with multilingual support (23 languages).

This is the only ComfyUI Chatterbox implementation that supports the multilingual model!

Features

  • Multilingual TTS: Support for 23 languages with language-specific text processing
  • Standard TTS: High-quality English text-to-speech
  • Voice Conversion: Convert voice timbre while preserving speech content
  • Voice Cloning: Clone voices from reference audio prompts
  • ComfyUI Integration: Full integration with ComfyUI's model management system

Supported Languages

Code Language Code Language Code Language
ar Arabic he Hebrew pl Polish
da Danish hi Hindi pt Portuguese
de German it Italian ru Russian
el Greek ja Japanese sv Swedish
en English ko Korean sw Swahili
es Spanish ms Malay tr Turkish
fi Finnish nl Dutch zh Chinese
fr French no Norwegian

Nodes

Chatterbox Multilingual TTS

Generate speech in 23 languages with voice cloning support.

Inputs:

  • text: Text to synthesize (max 300 characters)
  • language: Target language selection
  • exaggeration: Emotional expressiveness (0.25-2.0)
  • temperature: Sampling randomness (0.05-5.0)
  • cfg_weight: CFG guidance weight (0.0-1.0, set to 0 for language transfer)
  • audio_prompt: Optional reference audio for voice cloning

Chatterbox TTS

Standard English TTS with high quality output.

Chatterbox Voice Conversion

Convert the voice in an audio file to match a target voice.

Installation

Method 1: ComfyUI Manager (Recommended)

Search for "ComfyUI-For-ChatterBox" in ComfyUI Manager and install.

Method 2: Manual Installation

cd ComfyUI/custom_nodes
git clone https://github.com/your-repo/ComfyUI-For-ChatterBox.git
cd ComfyUI-For-ChatterBox
pip install -r requirements.txt

Model Downloads

Models are automatically downloaded from HuggingFace on first use:

  • Standard TTS: models/tts/chatterbox/resembleai_default_voice/
  • Multilingual TTS: models/tts/chatterbox/resembleai_multilingual/

Tips for Best Results

Language Transfer

When using a reference audio in a different language than the target:

  • Set cfg_weight to 0 to mitigate accent transfer
  • Use a reference audio that matches the target language for best quality

Japanese

Japanese text is automatically converted to hiragana (kanji → hiragana, katakana preserved).

Chinese

Chinese characters are converted to Cangjie codes for tokenization.

Korean

Korean syllables are decomposed into Jamo for proper pronunciation.

Optional Dependencies

  • resemble-perth: Audio watermarking support
  • dicta_onnx: Hebrew diacritization
  • russian_text_stresser: Russian stress marking

Credits

License

MIT License

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages