ComfyUI Extension
ComfyUI-ElevenLabs
API调用节点
GuardSkill/ComfyUI-ElevenLabs
Nodes2
On cloudLocal install
Categoryaudio/ElevenLabs
Stars0
Updated9 months ago
Readme
ComfyUI-ElevenLabs
ElevenLabs Text-to-Speech and Voice Query nodes for ComfyUI.

Features
- ElevenLabsVoiceQuery: Query ElevenLabs shared voice library with filters and random selection
- ElevenLabsTTS: Generate speech from text using ElevenLabs API
Installation
- Clone this repository into your ComfyUI custom_nodes folder:
cd ComfyUI/custom_nodes
git clone https://github.com/GuardSkill/ComfyUI-ElevenLabs.git
- Install dependencies:
cd ComfyUI-ElevenLabs
pip install -r requirements.txt
- Restart ComfyUI
Usage
ElevenLabsVoiceQuery
Query the ElevenLabs shared voice library to find and select voices based on various criteria.
Parameters:
api_key: Your ElevenLabs API keyseed: Random seed for reproducible voice selection (0 to 18446744073709551615)category: Voice category filter (all, high_quality, professional)gender: Gender filter (all, female, male, neutral)age: Age filter (all, middle-aged, old, young)language: Language code filter (all, ar, bg, cs, da, de, el, en, es, fi, etc.)locale: Locale filter (all, cmn-CN, cmn-TW, en-US, ja-JP, etc.)use_cases: Use case filter (all, advertisement, conversational, etc.)descriptive: Voice descriptive filter (all, calm, professional, warm, etc.)page_size: Number of voices to query (1-100)
Outputs:
voice_id: The selected voice ID
Features:
- All filters support "all" option to skip filtering on that criterion
- Random seed ensures reproducible voice selection
- Automatically retries API calls up to 3 times on failure
ElevenLabsTTS
Generate speech audio from text using a voice ID.
Parameters:
text: The text to convert to speechvoice_id: The ElevenLabs voice ID (can be connected from ElevenLabsVoiceQuery)model_id: The TTS model to use (eleven_multilingual_v2, eleven_turbo_v2, etc.)api_key: Your ElevenLabs API key
Outputs:
audio: Generated audio in ComfyUI AUDIO format
Features:
- Automatically retries API calls up to 3 times on failure
- Supports multiple ElevenLabs TTS models
Requirements
- elevenlabs
- librosa
- requests
License
MIT