This ComfyUI custom node suite provides a unified, multi-engine platform for advanced Text-to-Speech (TTS) and Voice Conversion (VC). It integrates numerous state-of-the-art engines, offering flexible, high-quality audio generation and manipulation capabilities for users within the ComfyUI ecosystem.

How It Works

The suite features a modular, universal streaming architecture, enabling seamless integration of diverse TTS (Chatterbox, F5-TTS, Higgs Audio 2, VibeVoice, IndexTTS-2) and VC (RVC, ChatterBox VC) engines. Key functionalities include advanced text processing with character/language switching, SRT timing synchronization, voice cloning, and experimental silent video analysis.

Quick Start & Requirements

• Primary install / run command: Recommended via ComfyUI Manager for one-click installation and dependency management. Manual installation involves cloning the repository and running an installer script or pip install -r requirements.txt.
• Non-default prerequisites and dependencies: ComfyUI installation, Python 3.12+, and specific system libraries for Linux/macOS (e.g., portaudio). Windows requires no additional system dependencies. Many engine models auto-download on first use; manual downloads are also supported.

Highlighted Details

• Extensive Engine Support: Integrates Chatterbox (23-lang official), F5-TTS, Higgs Audio 2 (high-fidelity cloning), VibeVoice (long-form generation), IndexTTS-2 (emotion control), and RVC (real-time VC).
• Advanced Features: Supports SRT timing, character/language switching syntax, pause tags, unlimited text length with smart chunking, and iterative voice conversion refinement.
• Voice Cloning: Higgs Audio 2 enables cloning from short audio samples.
• Emotion Control: IndexTTS-2 offers dynamic emotion expression via text analysis or direct control.

Maintenance & Community

Developed by diodiogod, evolving from ShmuelRonen's ChatterBox Voice project. Credits ResembleAI for ChatterboxTTS. The GitHub repository serves as the primary community hub.

Licensing & Compatibility

Released under the MIT License, permitting broad usage, including commercial applications. It functions as a ComfyUI custom node, with noted compatibility adjustments for Python 3.13.

Limitations & Caveats

The Silent Speech Analyzer and OpenSeeFace provider for mouth movement detection are experimental. F5-TTS requires adherence to specific best practices for optimal generation quality and to avoid inference failures. Large model downloads are necessary for many engines.