HeyGen create avatar from a text prompt API vs Sume

HeyGen creates an avatar from text with POST /v3/avatars and type: "prompt": the prompt holds up to 1000 characters and aspect_ratio takes 16:9, 9:16, 1:1, 4:5, 5:4 or auto. Sume has the same text-only path as the Prompt input on POST /v1/avatar-1.0/generate, as an async job you poll.

What does HeyGen's prompt avatar call take?

Per the page read 2026-10-01, you describe a character in prompt and the response is a look you poll until status leaves processing. Omitting aspect_ratio is the same as auto. The page warns that details you leave out get invented, including text on badges and props, so inspect the image before scripting around it.

How does Sume's Prompt input work?

The avatar docs list three ways to make an avatar: Prompt, Profile (props) and Image (photo). Prompt is for "an avatar from text only". Creation uses a top-level avatar_handle plus an input union; the handle may include a leading @ and is stored without it. Each request creates a job; poll it, then use the returned handle or resource id for avatar videos. The pages I read do not state a prompt length limit or an aspect ratio field for this route.

How do the fields line up?

What should I check before using the avatar?

Look at the generated result first, then use the handle in a talking video. If you need a likeness from a real person, use the photo input instead; avatar from prompt, props or photo compares the three inputs.