ElevenLabs Avatars has no API at launch: the Sume Avatar API

ElevenLabs Avatars does not offer an API at launch. The ElevenLabs documentation says an API for Avatars is planned, and the product is used through its interface on paid plans. If you must generate avatar videos from code or an agent today, Sume Avatar 1.0 exposes them as plain HTTP: POST /v1/avatar-1.0/talking-video returns a job you can poll or receive by webhook.

What ElevenLabs documents

Avatars is available on all paid plans, takes reference images or a text prompt, and can use any voice from the library, including cloned voices. The lip-sync model is selected automatically. Credits follow the Image & Video pricing and are shared across ElevenCreative tools.

The Sume route

An avatar video needs an avatar handle plus exactly one of script or video_inputs. The default communication mode is async and returns 202. Poll /v1/jobs/:id/status or /result, or subscribe a webhook that is signed with HMAC SHA256.

Send an idempotency key with the request so a retry cannot create a second paid job.

curl -X POST https://api.sume.com/v1/avatar-1.0/talking-video \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: launch-clip-001" \
  -d '{"avatar_handle":"product_host","quality":"standard","script":"Welcome. Here is the new feature in thirty seconds."}'

When the interface is enough

If a person is making a handful of clips and wants the ElevenLabs voice library, the interface may be all you need. The API matters when clips are made by a scheduler, a CMS or an agent, and you want admission, spend and retries to be explicit.

Re-check the ElevenLabs page before you plan around any future API; the vendor has not given a date.