Mascot dance ad: Kling motion control, orientation and sound
Drive a still mascot with a motion reference video through Kling 3.0 Motion Control on Sume. Settings that matter: duration, orientation and original sound.

To make a still brand mascot dance in an ad, send the still and a motion reference video to POST /v1/kling/3.0/motion-control. The reference drives all output motion, and your image supplies the character. Set duration_seconds from 1 to 30 to match the reference, decide on keep_original_sound, and choose character_orientation.
What drives what
The docs are explicit about roles. motion_video_url is required and drives all output motion. The visual source is exactly one of image_url, or a ready avatar by avatar_id or avatar_handle. The optional prompt steers appearance only, not the movement.
The provider has no duration knob. The length of the motion video sets the output length. So duration_seconds is the reservation basis: it tells Sume how much to hold for the job, and it has to describe your reference. If you declare 8 seconds and the reference runs 20, the reservation does not match the work, so measure the file before you submit.
The request
This is the shape from the Kling motion control doc, with your own URLs. Add Idempotency-Key, since all four submit paths accept it.
curl -X POST https://api.sume.com/v1/kling/3.0/motion-control \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: mascot-dance-001" \
-d '{
"avatar_handle": "reference_presenter",
"motion_video_url": "https://example.com/dance-reference.mp4",
"duration_seconds": 8,
"keep_original_sound": false,
"character_orientation": "video"
}'Sound and orientation
keep_original_sound defaults to true, which keeps the audio of your motion reference. Send false for a silent clip, which is what you want when the ad has its own music bed. character_orientation is video by default, so the character follows the facing of the reference. Use image to keep the facing of your still.
| Field | Values | Default |
|---|---|---|
| duration_seconds | 1 to 30 | required |
| keep_original_sound | true or false | true |
| character_orientation | video or image | video |
| prompt | Appearance steering only | optional |
Rights and review
A motion reference is someone's performance. Use footage you own or have the right to use, and check the result before it goes live. For clips over 30 seconds, split the reference into parts and join the outputs on a Timeline render. Poll the job on GET /v1/jobs/{id}/status or pass a webhook_url with mode: async, and store the Sume-hosted output URL, not a provider URL.
Sources
Related posts
More in Use cases
- Mattress Black Friday video ad: a dimmed bedroom clip and a price card
Make a mattress Black Friday ad with AI: one 8-second bedroom clip, a dim pass, and a price-card overlay on top, about $1.12 in Sume jobs.
- Meta ads on phones taller than 9:16: zoom or black bars, plan for both
Meta may zoom a 9:16 ad on taller phones or pad it with black. Keep text in the safe zone for both and test it with a free Sume crop check on a 1080x1920 clip.
- Meta's aspect ratios view fills missing variations: supply all four
If you upload fewer than four ratios, Meta pads, crops or AI-expands the rest. See which ratios Sume renders natively and how to crop the one it cannot.
- Meta catalog video: the ad spend per SKU that pays for the clip
Meta says catalog ad sets with product video saw 20% more conversions per dollar. Break-even spend per SKU is 5 x clip cost x CPA/margin; $0.93 and $2.19 clips.
Written by Sume