AI presenter for YouTube Shorts: format limits and Sume output

A YouTube Short can be square or vertical and up to three minutes long, and a single Sume avatar video job covers up to 60 seconds in 9:16 or 1:1, so one job fits inside a Short. For anything longer than a minute, render several jobs and join them.

What YouTube says

From YouTube's Shorts help page, read 2026-10-08:

What Sume gives you

Per Generate avatar video: aspect ratios 1:1, 3:4, 9:16, 4:3 and 16:9; default 9:16; 720p; estimated 4-60 seconds per job. Of those, 9:16 and 1:1 match the Shorts shapes. 3:4 is portrait but is not named by YouTube's page, so I would not rely on it for Shorts without testing.

• Under 60 s: one job, 9:16, captions on.
• 60 s to 3 min: two to three jobs with the same avatar_handle, joined in order.
• Background music: remember the Content ID rule above for anything over a minute.

Music and claims

The Content ID rule is the one that costs money. A Short over one minute with an active claim is blocked everywhere and loses monetization, per YouTube. If you add music to a long presenter Short, use a source you hold the rights to, or keep the Short at one minute. Sume's Music Router can generate a bed, but check the licensing terms of any track you use.

Also check the Audio Library limits: a song may be usable for 90 seconds in a three-minute Short, and some only for 60 or 30.

A plan

Write a 55-second script for a one-job Short, or split a longer script on topic boundaries. Preview first frames at 9:16, render with inline captions, check captions.status in each result, then upload. Duration windows and caption rules are on the avatar pages; platform rules change, so re-read YouTube's page before a launch.

Checking the finished Short

Upload as unlisted first and watch it on a phone. Confirm that YouTube classifies it as a Short, that captions are visible, and that no copyright claim appears on any music. Only then make it public.

Remember that the Shorts rules are YouTube's, and they change. The three-minute length and the vertical-or-square rule are as of the page read on 2026-10-08. Sume's own limits, 60 seconds per job and 720p, are the fixed part of this plan.