Discord
This is the real Kling 3.0 motion control workflow — reference video plus subject photo, not a text prompt. Kling 3.0 Motion Control takes the dance clip as the reference motion and the photo as the subject, then re-performs the exact same choreography. Everything below is the actual input and output from the AI Baby Dance examples.
Want a different subject or move? AI Baby Dance and AI Motion Control run the same reference-video motion control on other photos.
1 dance clip
the reference motion Kling 3.0 motion control follows
1 photo
the subject Kling 3.0 motion control puts in the dance
Same moves
Kling 3.0 copies the choreography exactly
3 welcome credits
for new accounts
Reference-driven dance following — the clip supplies the choreography, your photo supplies the star.
01
Upload a 3–10s clip of the dance you want as the reference video. Kling 3.0 motion control reads the moves, timing and framing from it.
02
Drop in one clear, front-facing photo with the body visible. Kling 3.0 motion control uses this person as the subject.
03
Kling 3.0 Motion Control re-performs the exact same dance with your subject, keeps the face recognisable, and hands you a short loop for Reels or TikTok.
One dance clip, one photo. Each line is a ready starting point for Kling 3.0 motion control.
Trend cover — feed a viral dance clip and let your photo copy it move for move.
Birthday clip — a friend or baby performing a fun dance for the group chat.
Announcement — a reveal set to a dance everyone already knows.
Duo post — run each photo against the same dance, cut them into a side-by-side.
Throwback — bring an old photo to life on a modern dance.
Cosplay cover — your character photo copying a choreography from the fandom.
Pet edition — swap in the family dog as the subject on the same dance.
Holiday card — a festive dance clip in a seasonal outfit.
Milestone — pair the clip with a caption about the moment.
Group remix — each friend photo on the same reference dance.
Side-by-side — show the dance clip next to your subject's version.
Weekly challenge — pick this week's dance and drop your photo into it.
Do
Don't
Kling 3.0 motion control is reference-driven video generation on Voor AI: you give a dance clip and a subject photo, and Kling 3.0 Motion Control re-performs that exact choreography with the person in the photo. The dance video supplies the moves and timing; your photo supplies the star. It is not a text prompt that invents random dancing.
Add a short dance clip as the reference video and a photo as the subject. Kling 3.0 reads the choreography, timing and framing from the clip and re-performs the identical dance with the person from your photo, keeping the face and outfit recognisable.
You need both. Kling 3.0 motion control does not invent choreography from text — it copies the motion from a real dance clip. Upload a 3–10s clip of the dance you want plus one clear photo, and the subject will perform that specific dance.
A steady 3–10s clip with the dancer's full body in frame, simple full-body moves and a static camera. Fast, shaky, heavily cut clips are harder to copy cleanly. The clearer the reference motion, the more faithful the subject's version.
A sharp, well-lit, front-facing photo with the upper body and ideally the legs visible. Kling 3.0 motion control needs to see the subject to place them into the dance, so avoid heavy shadow, blur or tight crops that hide the body.
That is the goal — Kling 3.0 motion control keeps the face, hair and outfit recognisable so it clearly reads as your subject, just dancing. If a result drifts, re-upload a sharper, front-facing photo and reroll.
New accounts receive 3 welcome credits. Additional Kling motion-control clips require purchased credits.
The motion-control collection on Voor AI preselects kwaivgi/kling-v3-motion-control — the same reference-video workflow shown in the examples above.
Both use the same Kling reference-video motion control on Voor AI. This page is branded for Kling 3.0 searches and shows the baby dance examples; AI Motion Control covers the same workflow with generic subject wording.
Yes — the same motion control works on any subject you photograph. Add the dance clip and a photo of the subject, and Kling 3.0 drops whatever you give it into the dance. For babies specifically, see the AI Baby Dance page.
Whichever Kling 3.0 motion control search brought you here, the workflow up top is the answer: one dance clip as the reference motion, one clear photo as the subject, and Kling 3.0 Motion Control re-performs the same dance — face kept recognisable, 3 welcome credits.
Add a dance clip, add one photo, and let Kling 3.0 motion control make the subject perform that exact choreography.
Generate with Kling 3.0 motion control