Comfyui-Orpheus

Comfyui-Orpheus
★ 10

语音合成情感语音多语言ComfyUI节点
Comfyui-Orpheus 在 ComfyUI 中提供高质量多语言（8种）情感TTS，支持24种声音并输出可用音频。
💡 在ComfyUI流程中将文本生成带情感的多语言语音
🍴 2 Forks💻 Python🔄 2025-04-16
📦
网盘下载
复制链接后前往夸克网盘下载
https://pan.quark.cn/s/af9fbf81e746
📦 requirements.txt
sounddevice
snac
📄 README

ComfyUI-Orpheus Node

2 custom nodes for ComfyUI that enables text-to-speech generation using the GGUF Orpheus model with emotional speech capabilities.

Features

  • High-quality text-to-speech synthesis
  • Multiple voice options (24 different voices, depend language used) in English, French, Spanish, Italian, Chinese, Korean, German, Hindi
  • Emotional speech capabilities
  • Seamless integration with ComfyUI workflow

    • Available Voices

  • English Voices:

    • supported tags : chuckle, cough, gasp, groan, laugh, sigh, sniffle, yawn

  • tara – Female voice
  • leah – Female voice
  • jess – Female voice
  • leo – Male voice
  • dan – Male voice
  • mia – Female voice
  • zac – Male voice
  • zoe – Female voice

  • French Voices:

    • supported tags : chuckle, cough, gasp, groan, laugh, sigh, sniffle, whimper, yawn

  • pierre – Male voice
  • amelie – Female voice
  • marie – Female voice (doesn’t works well)

  • German Voices:

    • supported tags : chuckle, cough, gasp, groan, laugh, sigh, sniffle, yawn

  • jana – Female voice
  • thomas – Male voice
  • max – Male voice

  • Korean Voices:

    • supported tags : 한숨, 헐, 헛기침, 훌쩍, 하품, 낄낄, 신음, 작은 웃음, 기침, 으르렁

  • 유나 – ?^^
  • 준서 – ?^^

  • Chinese Voices:

    • supported tags : 嬉笑, 轻笑, 呻吟, 大笑, 咳嗽, 抽鼻子, 咳

  • 长乐 – ?^^
  • 白芷 – ?^^

  • Hindi:

    • supported tags : unknow

  • ऋतिका – ? ^^

  • Spanish Voices:

    • supported tags : groan, chuckle, gasp, resoplido, laugh, yawn, cough

  • javi – Male voice
  • sergio – Male voice
  • maria – Female voice

  • Italian Voices:

    • supported tags : sigh, laugh, cough, sniffle, groan, yawn, gemito, gasp

  • pietro – Male voice
  • giulia – Female voice
  • carlo – Male voice

    • Requirements

  • Last ComfyUI version with python 3.12.9 (may be works with older versions but I haven’t test it)

    • Installation

  • Clone this repository into your ComfyUI custom nodes directory:

    • cd ComfyUI/custom_nodes
git clone https://github.com/your-repo/ComfyUI-Orpheus.git

  • Install the required dependencies:

    • load venv and :

    pip install -r ComfyUI-Orpheus/requirements.txt

    Or use python_embeded :

    python_embeded\python.exe -m pip install -r ComfyUI-Orpheus/requirements.txt

  • Download the required GGUF model from FreddyAboulton huggingface page, and place it in your ComfyUI models directory under models/unet/.
  • GPU Support

    • On windows, default installation of llama-cpp-python doesn’t take NVIDIA GPU support. If you want NVIDIA GPU Support you need to locate nvcc.exe folder and:

    set CMAKE_ARGS="-DGGML_CUDA=on"
set CUDA_CXX="YOUR_CUDA_DIR\v12.x.x\bin\nvcc.exe"
python_embeded\python.exe -m pip install llama-cpp-python[server] --upgrade --force-reinstall --no-cache-dir

    Be patient, can takes time…

    Usage

  • In ComfyUI, locate the “Orpheus ⛓️” node in the node menu.

  • Configure the node parameters:

  • model_name: Select your GGUF model
  • voice: Choose from available voices
  • prompt: Enter the text you want to convert to speech, you can add emotive tags :\, \, \, \, \, \, \, \

  • Connect the node outputs:
  • audio: Contains the generated audio waveform and sample rate

    • Limitations

  • Maximum text length determined by MAX_TOKENS
  • Processing speed depends on GPU capabilities
  • Requires CUDA support for optimal performance

    • Credits

  • Original Orpheus implementation
  • Freddy Aboulton