Creatify custom avatar lipsync_input and consent video, explained

Creatify's Custom Avatar API builds an avatar from your own videos: a lipsync_input MP4 used for lipsync training, plus a consent video, in video/mp4 or video/quicktime. Sume's avatar API takes a different route: a reference image sent as a photo input, which produces an avatar you reuse by handle.

Creatify details are from its Custom Avatar page, read 2026-10-01. Sume details are from Avatar and Avatar videos.

What does Creatify's custom avatar request need?

Two endpoints are documented. POST /api/personas_v2/ is a multipart upload of the files; POST /api/personas/ takes publicly accessible URLs as JSON. Both list lipsync_input, creator_name, gender and video_scene as required fields. The page describes uploading your own consent and lipsync videos, and then says to poll GET /api/personas/{id}/ for the approval status.

How does Sume create an avatar from an image?

Per the Avatar docs, use the image route when you have a reference image; the API input type is photo:

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-image-001" \
  -d '{
    "avatar_handle": "reference_presenter",
    "input": {
      "type": "photo",
      "image_url": "https://example.com/reference.png"
    }
  }'

How do I use the avatar afterwards?

Launch requests for Avatar Video reference a ready avatar with a top-level avatar_handle. The docs also say an avatar video uses one resolved avatar per final video. The image_url must be a fetchable public HTTPS URL. For multi-host layouts, see one avatar per video.

Which approach fits which input?

If what you have is footage of a person speaking, Creatify's page is the one that documents video input. If you have a still image, Sume's photo input starts from that image. Check each vendor's current terms for consent requirements before using a real person's likeness; this post is not legal advice.