Branded coach character repeats your phone-filmed move: $3.15 per 20 s
Film a move on your phone, give Kling motion control a still of your coach mascot, and the mascot repeats the move. 20 seconds bills $3.15 at $0.1575 a second.

The short answer
Kling motion control makes a still character repeat the movement in a reference video. Send your mascot image and a phone clip to POST /v1/kling/3.0/motion-control on Sume. Billing is $0.1575 per second of duration_seconds, rounded up, so a 20-second routine costs $3.15. The clip can run 1 to 30 seconds.
Why film it first
A fitness brand, a language app or a kids' channel often has a character and no animator. Motion control inverts the usual order: the movement comes from real footage, and the model repaints that movement onto your character. A stretching routine, a dance, a wave and a point-at-the-whiteboard all work the same way.
What to send
The request body has four fields that matter. image_url is the character (or use avatar_id or avatar_handle instead, never both). motion_video_url is your phone clip. duration_seconds is 1 to 30 and sets the reservation. character_orientation is video or image: choose video when the character should follow the clip's facing, image when it should keep the facing in your still.
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: coach-001" \
-d '{
"image_url": "https://example.com/coach.png",
"motion_video_url": "https://example.com/stretch.mp4",
"duration_seconds": 20,
"prompt": "Friendly cartoon coach in a teal tracksuit",
"keep_original_sound": false,
"character_orientation": "video"
}'Prompt and sound
The prompt describes appearance only, not the motion. The strict body rejects extra fields such as reference_*_urls. keep_original_sound defaults to true, so set it to false if your phone clip has room noise or your voice counting reps.
Cost of a routine
Price is a flat rate per second on ceil(duration_seconds). List is $0.126, and Sume bills list x 1.25.
| Routine length | Billed price |
|---|---|
| 5 s | $0.7875 |
| 10 s | $1.575 |
| 20 s | $3.15 |
| 30 s | $4.725 |
Filming tips
Film the clip with the whole body in frame, a plain background and the same framing as the character's still. Keep the take close to the length you reserve, because duration_seconds is the reservation basis. If the routine runs longer than 30 seconds, film it as two takes and join the results in a Timeline render.
Sources
Related posts
More in Use cases
- Budget a translated script per sentence before you dub it
Turn each source sentence's seconds into a character budget before translating, so the dubbed read fits the slot. A method with Sume sentence segments.
- Reel speed 0.5x to 3x: burn captions after the final speed change
Instagram lets creators change Reel speed from 0.5x to 3x. Burned-in captions speed up with the clip, so a 1.5 s line is 0.5 s at 3x. Burn them last.
- Call center roleplay training: customer persona clips with avatars
Record angry or confused customer prompts for agent practice with Sume avatars: one avatar per video, 10-second scripts, cost of a 24-clip scenario bank.
- Can a 30-second AI clip be a YouTube Short? Yes, up to 3 minutes
YouTube Help says a Short can be up to 3 minutes, so a 30-second AI clip fits. How to pick the model length and cut it to the file you upload, with Sume docs.
Written by Sume