YouTube avatar vs Sume Avatar: selfie capture or prompt and photo

A YouTube avatar is a digital version of you, built from a live selfie capture of your face and voice, which you can then place in videos made with YouTube's AI tools. To create one you must be 18 or older and the channel owner of an existing YouTube channel. Sume's Avatar 1.0 is different: you create an avatar from a text prompt, from structured traits, or from a reference image, give it a handle, and use that handle in talking videos by API.

The two are not interchangeable. A YouTube avatar lives in YouTube's apps; a Sume avatar is a resource in your workspace. This post sets the facts side by side.

How is a YouTube avatar created?

Per YouTube's help page, you create it once, in either the YouTube app or the YouTube Create app, and a change in one app shows in the other. In the YouTube app, tap Create, open AI playground from the top right, tap Create video, then Make a video with my avatar, and follow the prompts to record your selfie and voice. You preview it, then use it or retake it.

Its tips for a recording that is accepted include: phone at eye level, lighting that is not too dim or bright, eyes, nose and mouth visible with no sunglasses, masks or hats, a quiet room with no other voices, no other faces in the background, and the account owner doing the capture, never a minor.

Where can I use it, and how do I remove it?

Once you have an avatar, you add it during video creation: from AI playground, or on eligible Shorts in the feed through Remix, then Reimagine, then Add me to this scene. In YouTube Create you tap Generate video and pick your avatar as a reference in your prompt.

Avatars and videos are managed separately, the page says. Deleting your avatar permanently removes the recorded selfie and voice data and stops new content with it, but doesn't delete videos you already published or their remixes. Deleting a video doesn't delete the avatar. An avatar that isn't used to create new content for 3 years is deleted automatically.

How is a Sume avatar different?

Sume's Avatar 1.0 is created through POST /v1/avatar-1.0/generate with a top-level avatar_handle and an input of one of three types. Each request creates a job; when it completes, you use the avatar handle to make talking videos with POST /v1/avatar-1.0/talking-video.

What does a Sume request look like?

Create the avatar from a profile, wait for the job, then make the video. Scripts and multi-scene plans must estimate at 4 to 60 seconds, so a longer script is split into several jobs. Quality is standard, plus (default) or max.

curl -X POST https://api.sume.com/v1/avatar-1.0/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: avatar-profile-001" \
  -d '{
    "avatar_handle": "product_host",
    "input": {
      "type": "props",
      "ethnicity": "Asian",
      "sex": "female",
      "age": 28
    }
  }'

Which should I use?

If you want to appear in a Short as yourself with YouTube's own tools, the YouTube avatar is the route, and it needs you to be the channel owner. If you want a repeatable presenter across many videos, or a character that is not you, an avatar made through Sume's API is the route. Both end as video files you upload from Studio. For how YouTube treats AI likeness, see our post on likeness detection and AI avatars.