picturetovideo

Kling motion control · via API · Independent

Kling Motion Control: Make Any Photo Copy a Video’s Moves

Kling motion control copies the movement in a reference video onto the character in your photo — dances, gestures, walks, even the camera’s path. It is AI motion control in the literal sense: upload one image, pick a ready-made motion template, and choose the Motion tier (Kling Motion Control 2.6) for lower cost or Motion Pro (3.0) for sharper faces and hands. Clips run 3–30s, and you see the price before you render.

  • Motion Control 2.6 · 3.0
  • 3–30s
  • 720p · 1080p
  • 10 original templates
0/2500
Negative prompt
8 cr/s × 10s≈ $1.09Starter pack from $2.99
Character photo for Kling motion control: a young woman in a navy trouser suit standing against a white wall
01Your photo
Motion reference, pose 1: a man in a gold lamé outfit and cap dancing with his arms raised on a lawn
02Motion reference
Kling motion control illustration, pose 1: the woman from the photo doing the same arms-up dance in her navy suit
03Result · illustration
00:00 / 00:10

Illustrative keyframes · one timeline drives all three panels

01How it works

How AI motion control works

Kling motion control is not text to video with a dance prompt. It reads the body movement frame by frame from a reference clip and re-performs it with the person, pet or character in your still — keeping their face, outfit and proportions.
  1. 01Step 1/3

    A koi fish swims through an ink painting that comes alive

    + Push in+ Orbit+ Tracking

    Add one clear photo

    A full-body shot of one character works best: good light, arms and legs visible, framing similar to the motion clip. JPG, PNG or WebP.

  2. 02Step 2/3
    Studio8s1080p16:9Sound on152 cr

    Choose the motion

    Pick one of our original templates — hip-hop, ballet, salsa, robot pop and more — each performed by a fictional dancer in everyday clothes, with no music.

  3. 03Step 3/3
    Ready · 00:08Download MP4

    Pick 2.6 or 3.0, then render

    Motion (Kling Motion Control 2.6) is the budget choice; Motion Pro (3.0) holds faces and hands better. The clip is as long as the template, and credits are charged per second.

02Template library

Kling motion control templates

10 original moves, all safe for work and family feeds. Each card shows the 720p price from Motion (2.6) to Motion Pro (3.0). Select one and it loads into the generator and into the demo at the top of the page — then add your photo.

10 original moves · no music · everyday clothes

Original & silent

Every move is choreographed for PictureToVideo and performed by fictional dancers. No songs, no celebrity routines — nothing to get your clip muted.

Pick a move above — it loads straight into the generator.

03Which version?

Motion control AI: 2.6 vs 3.0

Same photo, same templates, two Kling motion control engines. Check the timing on 2.6, then re-render your favourite on 3.0 when the face is large in frame or has to hold through a turn.
Motion Control 2.6Motion Control 3.0
Best forDance trends, drafts, petsClose-ups, faces, hero clips
720p8 cr/s14 cr/s
1080p13 cr/s19 cr/s
10-second clip, 720p80 cr ≈ $1.09140 cr ≈ $1.90
Reference length3–30s3–30s
Faces & handsGood at full-body distanceSharper, steadier identity
ReleasedDec 2025Mar 2026
04Motion transfer vs prompts

Kling motion control or image to video?

Veo 3.1, Wan 2.5, Hailuo 02 and Kling 3.0 can all animate a photo here — but from a written prompt. Only the Kling Motion Control models take a reference clip, which is what makes the timing exact.
Kling Motion ControlImage to video (Kling · Veo · Wan · Hailuo)
Where the movement comes fromA reference clip, step for stepYour words — the model invents the motion
TimingFollows the clip beat by beatApproximate; varies from render to render
Clip length3–30sUp to 15s, depending on the model
Best forDances, choreographed gestures, walksCamera moves, natural motion, scenery
From8 cr/s4 cr/s
05Photo & clip tips

Photo and clip tips for a clean result

Show the whole body

If the reference dancer kicks, your character needs legs in frame. Crop from just above the head to just below the feet, and leave space on both sides for arm swings.

One character, plain background

The model maps one performer onto one character. Group photos and busy backgrounds make the model guess who should move — and what should stay still.

Match the framing

A full-body routine wants a full-body photo; a waist-up gesture wants a waist-up portrait. When the framing matches, Kling spends its effort on the motion instead of inventing missing limbs.

Don’t hide the joints

Long coats, crossed arms and hands in pockets hide the elbows, wrists and knees Kling tracks. Everyday clothes in a colour that contrasts with the background give the cleanest motion.

Light the face evenly

Soft, front-facing light keeps identity stable through turns. Hard shadows across the eyes are the most common reason a face drifts halfway through a dance.

Expect stylised physics

Cartoon characters and pets can follow human moves, but bodies with very different proportions are re-interpreted, not copied joint for joint. Keep their moves simple and short.

06Orientation

Image vs video orientation: 10s or 30s

Kling motion control lets either your photo or the reference clip decide which way the character faces. That one setting also caps how long the reference can be.
Character photo for Kling motion control: a young woman in a navy trouser suit standing against a white wall
Photo
Motion reference, pose 1: a man in a gold lamé outfit and cap dancing with his arms raised on a lawn
Sets facing
Reference

Max reference length

30s

Fits 10 of 10 templates

0s10s20s30s
Who decides
The reference clip sets the facing
Your character turns, faces and travels the way the reference dancer does, and camera moves inside the clip carry over.
Best for
Full routines, spins, walks toward camera
A max-length clip costs
240 cr on 2.6 (≈ $3.26) · 420 cr on 3.0 (≈ $5.71)

PictureToVideo templates run with video orientation, so every move in the library fits.

07Use cases

What people make with it

Social

Dance clips

Put yourself, a friend who said yes, or a mascot into an upbeat routine — no rehearsal, no second take.

Brands

Mascot moments

Make a 2D or 3D brand character wave, cheer or celebrate a launch while staying faithful to the original artwork.

Creators

Character blocking

Rough out a performance for an illustrated character before committing to hand animation or motion capture.

Learning

Movement demos

Show a warm-up routine or a folk step on an illustrated instructor, so the lesson travels without filming anyone.

08Cost per clip

What Kling motion control costs

You pay for the seconds you render. Our templates run 8–10 seconds, so most dances cost 64–80 credits on Kling Motion Control 2.6 at 720p.
ModelResolutionCredits / secWith audioTypical clip
Motion Control 2.6720p8—10s · 80 cr ≈ $1.09
Motion Control 2.61080p13—10s · 130 cr ≈ $1.77
Motion Control 3.0720p14—10s · 140 cr ≈ $1.90
Motion Control 3.01080p19—10s · 190 cr ≈ $2.58

Example: a 10-second Kling Motion Control 2.6 clip at 720p costs 80 credits (≈ $1.09 on the Pro plan). Failed renders are refunded automatically.

09Rules

Rules for photos of real people

  • Animate yourself, or people who have clearly agreed to it — consent first, every time.
  • No public figures, celebrities or lookalikes. Uploads must follow our content policy, and the video models apply their own safety filters.
  • Photos of children only if you are their parent or guardian, and only with wholesome, everyday moves.
  • No sexual, suggestive or mocking motion. The template library is family-friendly by design.
  • Label AI-generated clips when you post them, especially when a real person appears.
  • Breaking the content policy closes the account without a refund.
FAQ

Kling motion control FAQ

Still curious? Email support@picturetovideo.org — a human replies within a day.

01What is AI motion control, in plain words?
It is motion transfer. You give the model one still image of a character and one reference clip of someone moving, and it re-performs that movement with your character while keeping their face, clothes and proportions. Unlike text to video, you don’t describe the dance in words — the reference clip defines every step, gesture and beat.
02Is AI motion transfer the same thing as Kling motion control?
“Motion transfer” is the general name for the technique; Kling Motion Control is Kuaishou’s implementation of it, offered here in versions 2.6 and 3.0. Other motion-transfer tools exist, but on this site the Kling models are the only ones that accept a reference clip. Veo, Wan and Hailuo animate photos from a written prompt instead.
03How long can a Kling motion control clip be?
Kling accepts reference clips of 3 to 30 seconds when the character’s orientation follows the video, and up to 10 seconds when it follows the image. Seconds are rounded to whole numbers. Our templates run 8 to 10 seconds, and the clip you get back is as long as the template you choose.
04Should I choose Kling Motion Control 2.6 or 3.0?
Start with 2.6: it costs 8 credits per second at 720p and handles full-body dances well. Switch to 3.0, at 14 credits per second, when the face is large in frame or identity must hold through turns — it keeps features and hands steadier. Both accept the same photo and template, so re-rendering a winner takes one click.
05Is there a free motion control AI here?
Browsing templates and testing orientation is free, but renders are not: there is no free tier, so Kling motion control needs credits from a pack or plan. A 680-credit pack ($9.90) covers 8 ten-second Motion Control 2.6 clips at 720p, with failed renders refunded.
06What kind of photo gives the cleanest result?
One person, standing, photographed head to toe in even light against a simple background, with arms and legs clearly visible. Match the framing of the motion: full-body dances need full-body photos. Avoid sunglasses, hands in pockets, long coats and hard shadows across the face, and use at least 720 pixels on the short side.
07Can I upload my own dance video as the reference?
Not yet. Today every render on PictureToVideo uses one of our original templates, so you never copy someone else’s choreography or music by accident. Kling’s engine itself accepts any reference clip, which is why the rule matters wherever you use it: only use motion you recorded yourself or have the right to use.
08Will my Kling motion control video have sound?
No — and that is deliberate. Our templates are silent, so there are no song rights to worry about, and your result comes back silent too. Add a licensed track in CapCut, Premiere or your phone’s editor and trim it to the beat. Silent templates keep your clip free of music claims when you post it.
09Can I animate a photo of someone else?
Only with their clear permission. Public figures, celebrities and lookalikes are not allowed, and photos of children are allowed only from their parent or guardian. Prompts are screened, the video models apply their own safety filters, and accounts that animate people without consent are closed. When you post the clip, label it as AI-generated so nobody mistakes it for real footage.
10Is this Kling motion control tool run by Kuaishou?
No. PictureToVideo is an independent studio, not affiliated with, endorsed by or operated by Kuaishou or Kling AI. We access the Kling AI Motion Control models through their API and add our own templates, prompt screening and billing. Kling is a trademark of its owner; we use the name only to describe the models we run.
Keep exploring

Related tools & models

AI Dance Generator
AI Dance GeneratorFilter dances by person, pet or cartoon and check your photo.
Kling 2.6
Kling 2.6The model family behind Motion Control 2.6.
Photo to Video AI
Photo to Video AIAnimate a photo with a prompt instead of a dance.
Kling 3.0
Kling 3.0The 3.0 engine: multi-shot, 4K and steadier faces.

Your photo, their moves

Pick a template, add one clear photo and render in minutes. Credits are charged per second and refunded if a render fails.