Nodes/Elevenlabs-ComfyUI/ElevenLabs Voice Manager
ComfyUI Node

ElevenLabs Voice Manager

Inspect a voice without leaving the graph

By karthikg-09·Created 11 months ago·Updated 11 months ago· 1
ElevenLabs Voice Manager
    • voice_info
    api_key
    voice
    actionGet Info

    Every voice in your account has metadata hiding behind it - its ID, description, the settings it was created with, its labels. Normally you'd open the ElevenLabs dashboard to see any of it. This node reads it straight into your graph. Pick a voice, pick an action, get a formatted report.

    How it works

    You pick a voice from the dropdown (same cached "Name (voice_id)" list as the TTS node) and choose an action. The node calls the corresponding read-only endpoint:

    • Get Info - hits /v1/voices/{id} and reports name, ID, description, category, who created it, its stored settings (stability, similarity boost, style, speaker boost), and labels.
    • Get Settings - hits /v1/voices/{id}/settings and returns the voice's current settings block.

    Both come back as a formatted voice_info STRING, printed to the console too.

    Fair warning about the README: it describes list/delete actions for this node, but the actual node ships only these two inspect actions. There is no delete here. If you wanted to tidy up your voice library, that's still a dashboard job - this node is a viewer, not an editor.

    Inputs

    • api_key - your key.
    • voice - the dropdown of your account's voices.
    • action - Get Info (default) or Get Settings.

    Output: a voice_info STRING.

    Where it's actually useful

    Mostly as a debugging and planning tool. Two good uses: understanding why a voice sounds the way it does - check Get Info and you'll see the baked-in stability/similarity settings that voice was created with, which explains a lot when one voice behaves wildly differently from another in TTS. And pre-flight checks - before you push a cloned voice into production, verify its metadata and settings are what you think they are.

    Honestly, it's the least-exciting node in the pack, and that's fine. Utility nodes aren't supposed to be exciting. If you live in the TTS node all day and never touch this one, you're not missing much - but when you need to know what a voice is made of, it's faster than tabbing out to a browser.

    Install

    Same pack, same command:

    cd ComfyUI/custom_nodes
    git clone https://github.com/karthikg-09/Elevenlabs-ComfyUI.git
    cd ElevenLabs-ComfyUI && pip install -r requirements.txt
    

    or ComfyUI Manager → "Elevenlabs-ComfyUI" → restart. Add Node → ElevenLabs.

    CategoryElevenLabs

    Inputs (3)

    NameTypeDefaultDescription
    api_keySTRING
    voiceCOMBO21 options: Roger - Laid-Back, Casual, Resonant (CwhRBWXzGAHq8TQ4Fs17), Sarah - Mature, Reassuring, Confident (EXAVITQu4vr4xnSDxMaL), Laura - Enthusiast, Quirky Attitude (FGY2WhTYpPnrIDTdsKH5), Charlie - Deep, Confident, Energetic (IKne3meq5aSn9XLyUdCD), George - Warm, Captivating Storyteller (JBFqnCBsd6RMkjVDRZzb), Callum - Husky Trickster (N2lVS1w4EtoT3dr4eOWO), +15
    actionCOMBOGet Info2 options: Get Info, Get Settings

    Outputs (1)

    NameTypeDescription
    voice_infoSTRING