One avatar handle, three ratios: the same face in 9:16, 1:1, 16:9

To keep the same AI presenter across a vertical Short, a square feed post and a wide webinar slot, create one avatar and reuse its handle. On Sume's talking-video call the handle stays the same and only aspect_ratio changes. It supports 1:1, 3:4, 9:16, 4:3 and 16:9, with 9:16 as the default, and resolution is 720p at this time.

Make the avatar once

POST /v1/avatar-1.0/generate creates a reusable avatar from a prompt, structured traits (props) or a reference photo (photo). The create job is polled like any other, and the handle can be written with a leading @, which Sume stores without it. The create page has the request shapes. Creation is a flat price per avatar, so it is a one-time cost, not part of each clip.

Call it three times

Each ratio is its own job with its own Idempotency-Key. Keep the script identical and change one field.

What stays the same, and what you should check

The identity comes from the avatar, so the face is the same one you created. Check each ratio rather than assume it matches: approve each ratio on a first-frame preview before you pay for the full render: previews accept the same body, including aspect_ratio. A face that reads well in a tall frame can sit differently in a wide one.