AI avatar Instagram account: Avatar 1.0 clips, then label the profile

A workflow for an Instagram account fronted by a generated character: create it with Avatar 1.0, render 9:16 English clips, then add the AI profile label.

5 min readSume
All posts

Create the avatar once, render 9:16 clips from it, and label the profile before the first post. Social Media Today reports that Instagram's profile label now reads "AI generated profile" and that it limits the reach of profiles with AI-generated persons that do not label (read 2026-10-02). Sume handles the first two steps; the label is set in Instagram.

How do you create the avatar?

Create new avatar takes a prompt, a structured profile or a reference image to POST /v1/avatar-1.0/generate, with an avatar_handle. Poll the job, then use the returned handle for videos. Prefer a prompt or profile start over a photo of a real person.

How do you render a clip?

Call POST /v1/avatar-1.0/talking-video with the handle and a script. The default aspect_ratio is 9:16, resolution is 720p, and the script must plan to 4 to 60 seconds. Avatar 1.0 is English only today.

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: persona-clip-001" \
  -d '{"avatar_handle": "my_persona", "aspect_ratio": "9:16",
       "quality": "plus", "script": "Three things to check before you buy a tent."}'

What goes on the checklist before posting?

Run the same short list on every clip.

  • Profile has the AI generated profile label set in Instagram
  • Script is a product or topic explainer, not a first-person claim of use
  • Captions burned in if the clip will play muted
  • Job id and post URL logged in your ledger

Does the label cost reach?

The report says unlabeled AI personas lose reach. It does not say what labeled ones get. Do not assume the label is free or costly; watch your own numbers.

How do you keep the clips consistent?

Reuse the same avatar_handle and a fixed scene direction. Avatar videos accept a scene with a prompt or a photo, so you can keep one background across a series. Use quality of plus, the default, for routine posts, and max for a hero clip where turnaround matters less.

If a post will play muted, turn on inline captions in the same call. Inline captions burn after generation and do not create a separate billed caption job, per the docs.

What if you need a second language?

Avatar 1.0 speaks English only, so an account that posts in Spanish or Korean needs another method for the spoken part. Captions can carry on-screen text in other languages over a clip, but the avatar voice stays English.

Sources

Related posts

More in Use cases

All Use cases posts

Written by Sume