ElevenLabs Voice Retriever (Soze)

A small lookup utility, not a text-to-speech node. ElevenLabs is the well-known hosted voice-cloning and TTS API - every voice in your library, whether it's one of their stock voices or one you've cloned yourself, has a voice_id you can copy from their dashboard. This node takes that ID and pulls back the voice's data plus a sample audio file, so you can confirm you've got the right voice before wiring that ID into whatever else in your pipeline actually needs it.

It's a small integration in a pack whose other API-backed nodes lean heavily toward ComfyDeploy and fal.ai - this is the pack's only ElevenLabs touchpoint. If your real goal is text-to-speech inside ComfyUI, you'll want a dedicated ElevenLabs TTS pack alongside this one; treat this node as the "make sure I've got the right voice ID before I commit to it elsewhere" step, not the generation step itself.

How it works

You give it a voice_id, it calls ElevenLabs' API for that voice's details, and writes a preview audio sample to disk at the path you specify. It's genuinely just that - a lookup and a download, not generation. If your goal is producing new speech audio from text, this particular node isn't it; this pack's README lists exactly one ElevenLabs integration, and this is it.

The inputs and outputs that matter

• voice_id - the ID of an ElevenLabs voice, copied from your ElevenLabs dashboard or API.
• sample_save_path - where to save a preview audio clip of that voice.
• Outputs: voice_data (a STRING - the voice's metadata as returned by ElevenLabs) and status.

How to install it

Via ComfyUI Manager: search "Quality of Life Nodes for ComfyUI", install, restart. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/SozeInc/ComfyUI_Soze.git
pip install -r ComfyUI_Soze/requirements.txt

Needs ELEVENLABS_API_KEY set in your environment (or a .env file at the repo root, per the README) - an ElevenLabs account and API key, separate from anything ComfyDeploy- or fal-related elsewhere in this pack.

Common issues & troubleshooting

Auth error or nothing comes back. Check ELEVENLABS_API_KEY is set - same invisible-until-you-check-it failure mode as every other API-key node in this pack.

Wrong or empty voice_id. ElevenLabs voice IDs are account-specific - a voice ID from someone else's library, or a stock voice ID typo'd slightly, will fail to resolve. Copy it directly from the ElevenLabs dashboard rather than from memory.

You wanted generated speech, not a lookup. This node retrieves an existing voice's info and a short sample - it doesn't take arbitrary text and turn it into audio. If that's what you're after, you'll need a different node or a direct API call; it isn't in this pack.

sample_save_path folder doesn't exist. Like most file-writing nodes, this one expects the destination directory to already exist - create it first if you're pointing somewhere new.

You're not sure what's actually in voice_data. The schema types it as a plain STRING rather than a structured object, so expect it to be a text/JSON dump of whatever ElevenLabs' voice-details endpoint returns - name, category, and settings, most likely. Feed it into a JSON-parsing node from this same pack if you need to pull out a specific field rather than the whole blob.