ComfyUI-Chatterbox

[![Stargazers][stars-shield]][stars-url] [![Issues][issues-shield]][issues-url] [![Contributors][contributors-shield]][contributors-url] [![Forks][forks-shield]][forks-url]

About The Project

ComfyUI custom nodes for the powerful Resemble AI Chatterbox library. It enables seamless in-workflow Text-to-Speech and Voice Conversion, complete with deep integration into ComfyUI's model management system for efficient VRAM usage.

Major Update Notice

[!NOTE]

• 1.2.0: This version has been deeply refactored for better performance, stability, and alignment with the ComfyUI codebase. All parameters have been unlocked.

Features

• Long generation: No longer limited to 40 seconds.
• Chatterbox TTS Node: Synthesize speech from text with optional voice cloning from an audio prompt.
• Chatterbox Voice Conversion Node: Convert the voice in a source audio file to a target voice.
• Automatic Model Downloading: Models are automatically downloaded from Hugging Face on first use.
• Efficient VRAM Management: Full integration with ComfyUI's model patcher system to load models to GPU only when needed and offload them afterward.
• Detailed Generation Control: Fine-tune your audio output with parameters for speed, expressiveness, creativity, and quality.
• Accurate Progress Bars: Both console and UI progress bars reflect the true step-by-step generation process.

Getting Started

Installation

1. Install via ComfyUI Manager (Recommended):

   • Search for ComfyUI-Chatterbox in the ComfyUI Manager and install it.

2. Manual Installation:

   • Clone this repository into your ComfyUI/custom_nodes/ directory:

     git clone https://github.com/wildminder/ComfyUI-Chatterbox.git ComfyUI/custom_nodes/ComfyUI-Chatterbox
     

3. Install Dependencies:

   • Navigate to the new directory and install the required packages:

     cd ComfyUI/custom_nodes/ComfyUI-Chatterbox
     pip install -r r