AI avatar for YouTube videos: Shorts and long-form

To use an AI avatar in YouTube videos, create the avatar once, have it read your script, and render each video in the shape of its format: vertical 9:16 for Shorts and 16:9 for regular long-form videos. YouTube treats square or vertical uploads of up to 3 minutes as Shorts, so a long-form video needs the wider frame. With Sume, a Short can be one 9:16 talking video of up to 60 seconds, and a long-form video is 16:9 segments joined into one file of up to 30 minutes.

YouTube's rules are quoted from Upload YouTube Shorts and Understand three-minute YouTube Shorts, read on 2026-09-27. Sume facts come from Generate avatar video, Timeline 1.0, and the Sume API reference; anything called current behavior is read from Sume's code.

How do I create an AI avatar for my YouTube channel?

Create the presenter once with POST /v1/avatar-1.0/generate, from a text prompt, a profile of traits, or a reference photo, and give it a stable avatar_handle to send with every video. The docs recommend a stable handle so your app reuses a simple name instead of a generated id. Creating an avatar costs $0.95 per avatar.

How do I make a YouTube Short with an AI avatar?

Send one POST /v1/avatar-1.0/talking-video with aspect_ratio: "9:16", which is also the default. Sume accepts a script it estimates at 4–60 seconds; today the estimate counts 2.8 words per second, rounded up clip by clip, so a 60-second Short holds roughly 165 words at most.

• For a Short between 60 seconds and YouTube's 3 minutes, render up to three 9:16 segments and join them in one Timeline render of 180 seconds or less, over their own detached audio as in the long-form steps below. Timeline's default 1080×1920 frame is already vertical.
• Completed results can include public media.sume.com video artifacts; uploading the MP4 to YouTube is up to you. YouTube Shorts video specs compares the file with YouTube's settings.

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: channel-short-0927" \
  -d '{
    "avatar_handle": "channel_host",
    "script": "Three budgeting mistakes almost everyone makes, and one easy fix for each.",
    "aspect_ratio": "9:16",
    "quality": "plus"
  }'

How do I make a long-form YouTube video with an AI avatar?

Long-form is one POST /v1/timeline-1.0/render of up to 1,800 seconds. Its default frame is 1080×1920, so set output.width to 1920 and output.height to 1080. Build the footage one of two ways:

• Avatar segments: 16:9 talking videos of up to 60 seconds each, with the same avatar_handle, scene, and quality, joined over their own detached audio. How to make an AI avatar video longer than 60 seconds shows the request.
• Narration plus lip sync: speak the script with POST /v1/tts-1.0/generate in the avatar's voice (its avatar_handle, once the voice is ready), split the audio into parts, then lip sync a still to each part with POST /v1/veed/fabric-1.0, up to 300 seconds and 10 MB of Sume-hosted audio per job. Lip sync a long video walks through the split.
• A slot whose clip has another shape is placed by its fit: cover (the default), contain, stretch, or blur.

What are the limits?

Plan each video around these caps:

• English speech only on avatar talking videos, in current code: the clip prompt asks for English, and avatar voices are cloned in English. For another language, use the lip-sync route and set TTS's language.
• Captions go on in pieces. Today a caption job refuses a source longer than 60 seconds, so caption each avatar segment before the join; a lip-sync clip longer than that can't be captioned in one job.
• resolution is currently 720p on avatar videos; Timeline sets the frame of the joined file.

What does an AI avatar YouTube video cost?

Avatar video bills per second by quality tier: $0.184/s standard, $0.245/s plus, $0.55/s max (no product image). At the default plus tier, a 60-second Short is $14.70 and 10 minutes of long-form avatar footage is $147.00. The join is listed at $0.10 per output minute. On the lip-sync route, narration is $0.0475 per 1,000 characters and Fabric is $0.1875 per audio second (720p). Each is plus a 5.5% agent fee by default.