How to find an ElevenLabs voice ID
Updated Invalid Date
The picks
Press play to hear each voice read this guide's sample line.
These picks compare my fifteen voices. I own them, and ElevenLabs may pay me when you use them.
Dexter·Product instructions
Support delivery suits setup steps and error messages. Promotional excitement is outside its best fit.
Pulse·Short recap pipelines
Loose social-news pacing fits a quick update. Avoid grave news that needs a steadier delivery.
Derek·Scheduled announcements
Broadcast pacing fits names and dates that must stay clear. It can sound formal in a creator clip.
To find an ElevenLabs voice ID in the app, open My Voices, click the three-dot More actions menu on the voice and choose Copy voice ID. Through the API, call the list voices endpoint and read the voice_id field on each result. For my fifteen voices you can skip both routes. Every voice page on this site prints the ID, 20 characters of mixed case and digits, with a copy button beside it.
How do I copy a voice ID in the ElevenLabs app?
- Open My Voices.
- Find the voice and click More actions, the three dots.
- Choose Copy voice ID. It’s now on your clipboard.
A voice only shows up in My Voices once it’s on your account. For a Voice Library voice, add it first with the + button in the library, or open its share link while signed in, and it lands in My Voices on its own.
The Voice Library search box takes voice IDs too. That’s handy backwards. Paste an ID someone sent you and you’ll see which voice it belongs to before you build anything on it.
How do I get a voice ID through the ElevenLabs API?
Every request carries your API key in the xi-api-key header.
Voices on your account. GET https://api.elevenlabs.io/v2/voices returns the voices available to you, each with voice_id and name. You can narrow it with search, or with voice_type set to a value such as personal or community. Results come in pages of up to 100.
curl "https://api.elevenlabs.io/v2/voices?search=Dexter&page_size=10" \
-H "xi-api-key: $ELEVENLABS_API_KEY"
Voices in the shared library. GET https://api.elevenlabs.io/v1/shared-voices searches the Voice Library itself. Each result includes voice_id, name and public_owner_id, along with a free_users_allowed flag that deserves a look before you commit to a voice. It pages up to 100 at a time as well.
Adding a library voice from code. Send POST https://api.elevenlabs.io/v1/voices/add/{public_user_id}/{voice_id} with a JSON body that includes new_name. The response contains a voice_id. Store the one it returns and use that from then on.
With the ID in hand, it goes in the path of the speech request, POST /v1/text-to-speech/{voice_id}. Set model_id in the body while you’re there. If you leave it out, the endpoint falls back to eleven_multilingual_v2, and Eleven v4 has to be requested by name. My implementation guide has a complete working request.
Why does an ElevenLabs voice ID stop working?
Usually one of five things.
- A free plan with a library voice. ElevenLabs’ docs say Voice Library voices aren’t available through the API on the free tier. Copying the ID works. Generating with it from code takes a paid plan. The app plays by different rules, and there ElevenLabs only blocks free accounts from a library voice when the owner excluded free users or the voice carries a credit multiplier.
- Saving the name. When you add a library voice, you choose what to call it, so one person’s “Narrator” is another person’s “Dex” and both point at the same voice. Names are labels. Your config should hold the ID.
- A partial copy. Every ID on my voice pages is 20 characters of mixed upper case, lower case and digits. If yours is shorter, or you typed it in by hand, copy it again.
- A voice that stops being shared. Owners can withdraw a voice. With a notice period, it stays usable for people who saved or used it during that period, and the Get Voice response includes
disable_at_unixso your code can see the date coming. Without one, access ends at once. ElevenLabs lets you filter the library by notice period, and I’d do that for any project meant to run a long time. - Default voices. ElevenLabs says all its Default voices expire on December 31, 2026. Code that hardcodes a Default voice ID needs a library or cloned voice swapped in before that date.
Where should I keep an ElevenLabs voice ID?
Next to the model and the settings you tested, in the same file as the script. Trust me on this one. Six months on, “the deep one we used for the launch” means nothing to the API, and the voice list on your account will not look the way it does today.
Which of my voices suit API projects?
Dexter fits support flows and onboarding inside a product. It’s steady enough for step-by-step instructions and doesn’t rush the moment a user is supposed to act. Clarity is the whole design, which is why hype copy comes out polite in Dexter’s voice, so keep your launch announcement away from it. Its ID is Smxkoz0xiOoHo5WcSskf, and the full page is at Dexter.
Pulse suits automated recap and trending-story pipelines, where a script arrives from upstream and should sound like a friend catching you up. It’s loose and quick. Heavy news is where that looseness hurts, so test it on a short recap before trusting it with anything grim. Its ID is mhgBlD8CmCSdwLDOIJpA. See Pulse for news recaps and trends.
I’d use Dexter for product steps, Pulse for a quick recap, or Derek for a scheduled announcement. Their pages have the exact IDs; copying a name into an API voice-ID field won’t select it.
Sources
- ElevenLabs Help Center: How do I find the voice ID of my voices via the website and API?
- ElevenLabs API reference: List voices
- ElevenLabs API reference: List shared voices
- ElevenLabs API reference: Add shared voice
- ElevenLabs API reference: Create speech
- ElevenLabs API reference: Authentication
- ElevenLabs docs: Voice Library
- ElevenLabs docs: Voices
Related guides
Other jobs where I use the same voices.