Extensions/ComfyUI-KokoroTTS-Zh
ComfyUI Extension

ComfyUI-KokoroTTS-Zh

Kokoro TTS's Chinese comfyui node, 100 speakers

By ruanjianlun·Created about a year ago·Updated about a year ago· 2
ruanjianlun/ComfyUI-KokoroTTS-Zh
Nodes4
On cloudLocal install
CategoryKokoroTTS
Stars2
Updatedabout a year ago

Nodes (4)

KokoroModelLoader
KokoroTTS
KokoroTextInput
KokoroTTS
KokoroTTSGenerator
KokoroTTS
KokoroVoiceSelector
KokoroTTS
Readme

ComfyUI-KokoroTTS-Zh

A custom node extension for ComfyUI that integrates KokoroTTS for high-quality Chinese text-to-speech synthesis.

Introduction

ComfyUI-KokoroTTS-Zh is a collection of extension nodes for ComfyUI that integrates the KokoroTTS voice synthesis engine into ComfyUI workflows. It provides high-quality Chinese speech synthesis with support for multiple voices and speech speed adjustment.

Features

  • Load KokoroTTS models
  • Process multi-paragraph text input
  • Support for nearly 100 different Chinese voices (50 female and 50 male)
  • Control silence between paragraphs
  • Sentence joining option
  • Support for automatic and fixed speed modes
  • Generate high-quality 24kHz WAV audio
  • Compatible with ComfyUI audio nodes

Installation

Installation Steps

  1. Make sure ComfyUI is installed and running properly

  2. Clone this repository to the custom_nodes directory of ComfyUI:

    cd /path/to/ComfyUI/custom_nodes
git clone https://github.com/ruanjianlun/ComfyUI-KokoroTTS-Zh.git
    pathto\python_embeded\python.exe -m pip install "kokoro>=0.8.1" "misaki[zh]>=0.8.1"  "soundfile>=0.10.0"

  3. Install dependencies:

    cd ComfyUI-KokoroTTS-Zh
pip install -r requirements.txt

  4. Restart the ComfyUI service

Usage

In ComfyUI, you can use the following nodes:

  1. KokoroModelLoader: Loads the KokoroTTS model

    • Input: Hugging Face repository ID
    • Output: Model and voice pipeline objects

  2. KokoroTextInput: Input text to synthesize

    • Input: Multi-line text (split by paragraphs)
    • Output: Processed text paragraph list

  3. KokoroVoiceSelector: Select voice and silence settings

    • Input: Voice ID, silence length between paragraphs, whether to join sentences
    • Output: Voice settings dictionary

  4. KokoroTTSGenerator: Generate audio

    • Input: Model, pipeline, text, voice settings, output filename, speed mode, fixed speed value
    • Output: Audio object and file path

Basic Workflow

  1. Load the model
  2. Input text to synthesize
  3. Configure voice and silence settings
  4. Generate and save audio

Example

Here's an example of a basic workflow configuration:

  1. KokoroModelLoader → Load default model "hexgrad/Kokoro-82M-v1.1-zh"
  2. KokoroTextInput → Input "欢迎使用KokoroTTS！这是一个很棒的语音合成引擎。"
  3. KokoroVoiceSelector → Select female voice "zf_018", paragraph interval 5000, don't join sentences
  4. KokoroTTSGenerator → Generate audio, using automatic speed

Advanced Options

  • Speed Control: You can select "auto" or "fixed" speed mode in the KokoroTTSGenerator node

    • Auto mode: Dynamically adjusts speed based on text length to avoid fast speech with long text
    • Fixed mode: Uses the specified fixed speed value (0.5-2.0)

  • Sentence Joining: Choose whether to join multiple sentences in a paragraph for processing

License

MIT License

Contributing

Pull requests and issues are welcome.

Acknowledgements

  • KokoroTTS - High-quality Chinese speech synthesis engine
  • ComfyUI - Powerful node-based UI framework