−50%
Yearly PlansLimited-time yearly offer
Claim now →
Open AI Image
Image StudioPricing−50%Save up to 50% with annual billingAPIBlog
Open AI Image

Explore GPT Image 2.6, compare the GPT Image family, and create with the image models available on Open AI Image.

Open AI Image is independent and is not affiliated with or endorsed by OpenAI.

Open AI Image

  • AI Image Generator
  • AI Canvas
  • Image Upscaler
  • Pricing

GPT Image

  • GPT Image 2.6
  • GPT Image 2.5
  • GPT Image 2
  • Compare models

Resources

  • Prompt Library
  • FAQ
  • Help Center
  • Blog
  • API

More tools

  • Text to Video
  • Image to Video
  • Video Upscaler
English中文

© 2026 • Open AI Image All rights reserved.

support@openaiimage.com

AboutTerms of ServicePrivacy PolicyRefund Policy
Open AI Image
Built withShipAny

Motion Control AIMove Any Character with a Reference Video

Attach a still of your character and a 3–30 second clip of someone doing the movement. Kling Motion Control performs the same moves as your character. Runs on Kling 3.0 Motion Control and Kling 2.6 Motion Control.

Home / Tools / Motion Control

Limits and prices read from the live generator on October 31, 2026. The background clip at the top is Kling 3.0 showcase footage, not a motion-control output.

Motion control, also called motion transfer, takes two inputs: a picture that decides who is on screen and a video that decides what they do. The model keeps the look of your character (face, outfit, proportions) and replays the timing, poses and gestures from the clip, so a dance, a gesture or a blocked-out scene you filmed on a phone comes back performed by a mascot, an illustration or a rendered character.

Each task uses the current generator quote for its model, duration, resolution and references. Identical settings cost the same credits on every plan; subscriptions do not reduce the task credit charge.

Open motion control in the generatorSee plans and credit packs

How to use motion control

  1. 1

    Pick a motion control model

    In the composer at the top of this page (or the full generator), open the model menu and choose Kling 3.0 Motion Control or Kling 2.6 Motion Control. The page opens with Kling 3.0 Motion Control selected.

  2. 2

    Add the character image

    Upload one JPG or PNG of up to 10 MB, at least 300 px on each side, aspect ratio between 1:2.5 and 2.5:1. The upper body or full body should be clearly visible and not cropped at the joints that will move.

  3. 3

    Add the motion video and set orientation

    Attach one MP4 or MOV of 3 to 30 seconds, up to 100 MB: a single continuous shot of one performer works best. Set Character Orientation to image (up to 10 s) or video (up to 30 s).

  4. 4

    Optionally add a prompt, pick 720p or 1080p, generate

    The prompt is optional (up to 2,500 characters) and is for clothing, background or camera details; the movement comes from the video. The composer will not submit until both files are attached and shows the credit cost first.

Character orientation: image or video

Image (default)
The character keeps the facing and framing of your still. Best for a performer who stays front-on, such as gestures, talking and most dance moves. The motion video can be 3 to 10 seconds.
Video
The character follows the turns and direction changes in the clip. Use it when the performer spins, walks across the frame or turns away. The motion video can be 3 to 30 seconds.

Motion control models on Open AI Image

Each task uses the current generator quote for its model, duration, resolution and references. Identical settings cost the same credits on every plan; subscriptions do not reduce the task credit charge.

No model quotes are currently available. Please try again later.

What one motion-control clip costs

A job is billed on the length of your motion video, since the result is as long as the clip. The table shows the credits the generator charges on each plan today (plan discounts included, as a new subscriber) and their dollar value at the plan’s price per credit. Taxes not included; failed runs are not charged.

PlanBilled amountCreditsPrice per credit
Lite 300$14.90/month300/month$0.0497
Lite 300$156.00/year300/month$0.0433
Pro 600$29.00/month600/month$0.0483
Pro 600$174.00/year600/month$0.0242
Pro 1300$59.00/month1300/month$0.0454
Pro 1300$598.80/year1300/month$0.0384
Pro 2000$89.00/month2000/month$0.0445
Pro 2000$828.00/year2000/month$0.0345
Premium 3500$149.00/month3500/month$0.0426
Premium 3500$1188.00/year3500/month$0.0283
Premium 5200$219.00/month5200/month$0.0421
Premium 5200$2028.00/year5200/month$0.0325
Premium 7200$299.00/month7200/month$0.0415
Premium 7200$2748.00/year7200/month$0.0318
Credit pack$29.90350$0.0854
Credit pack$49.90600$0.0832
Credit pack$79.901000$0.0799
Credit pack$199.002500$0.0796
Credit pack$999.0013000$0.0768

Annual plans are billed yearly. Starter credits are valid for 180 days; other standard packs for 365 days. Taxes and personal offers are shown in checkout.

Eligible new verified accounts may claim up to 5 welcome credits for 7 days after human verification, while the promotion budget is available. This can cover a low-cost image, not a guaranteed free video. New video generation requires sufficient purchased credits or eligible bounded unlimited capacity. See pricing for every plan and pack, or the AI video price index for the per-second rate of every model.

Motion control specs

Models
Kling 3.0 Motion Control and Kling 2.6 Motion Control (Kuaishou)
Required inputs
Exactly one character image and one motion video; both are required
Character image
JPG or PNG, up to 10 MB, at least 300 px per side, aspect ratio 1:2.5 to 2.5:1
Motion video
MP4 or MOV, 3–30 seconds (3–10 s with orientation set to image), up to 100 MB
Output length
Follows the motion video; very fast or complex motion can come back slightly shorter
Resolution
720p or 1080p
Prompt
Optional, up to 2,500 characters; describes look, setting or camera, not the movement
Sound
The motion video’s original sound is kept; no new audio is generated
Billing
Per second of the motion video; failed runs are not charged
Watermark
None, on every plan
Generation time
Longer than most models: the completed runs in our job log took about 8 to 20 minutes

Tips for a clean motion transfer

  • Use a motion video with one performer, a steady camera and no cuts. Several people, hard cuts or a moving camera make it unclear whose movement to copy.
  • Match the framing: a full-body dance needs a full-body character image, and a waist-up gesture clip works best with a waist-up still.
  • Keep hands, arms and legs visible in the image. Limbs hidden behind the body or cropped by the frame have nothing to move.
  • Trim the motion video to the part you need before uploading. You are billed per second of the clip, and the result is as long as the clip.
  • Leave orientation on image for front-facing performances; switch to video only when the performer turns, because image mode caps the clip at 10 seconds.
  • Stylized, illustrated or clearly fictional characters are the reliable choice. Real people’s faces and copyrighted characters can be refused by the provider’s content rules; a refused run is not charged.

When to use another tool

You want to change something inside an existing video instead
Video to video. Replace a person or an object, change the background or restyle the take while its own motion stays.
You have the character but no motion clip
Image to video. The image becomes the first frame and the prompt describes the movement, up to 30 seconds on Seedance 2.5.
The same character should appear in a new scene from several pictures
Reference to video. Up to 30 reference images on Seedance 2.5 keep a character, product or set consistent across shots.
A finished clip ends one beat too early
AI video extender. Seedance 2.5 continues a 2–30 second clip from its last frame.

Motion control AI FAQ

What is motion control in AI video?
Motion control (motion transfer) generates a video in which the character from your image performs the movement from a reference video. The image decides who appears, the video decides what they do: poses, timing, gestures and dance steps. On Open AI Image it runs on Kling 3.0 Motion Control and Kling 2.6 Motion Control.
How do I use motion control on Open AI Image?
Choose Kling 3.0 or Kling 2.6 Motion Control in the generator, upload a character image and a 3–30 second motion video, set Character Orientation to image or video, optionally add a prompt for clothing or background, pick 720p or 1080p and generate. The composer shows the credit cost before you submit.
Is motion control AI free?
Eligible new verified accounts may claim up to 5 welcome credits for 7 days after human verification, while the promotion budget is available. This can cover a low-cost image, not a guaranteed free video. New video generation requires sufficient purchased credits or eligible bounded unlimited capacity.
How much does motion control cost?
Each task uses the current generator quote for its model, duration, resolution and references. Identical settings cost the same credits on every plan; subscriptions do not reduce the task credit charge.
How long can a motion control video be?
As long as your motion video: 3 to 30 seconds with Character Orientation set to video, or 3 to 10 seconds set to image. Very fast or complex motion can come back slightly shorter than the source.
Kling 3.0 or Kling 2.6 Motion Control: which should I use?
Each task uses the current generator quote for its model, duration, resolution and references. Identical settings cost the same credits on every plan; subscriptions do not reduce the task credit charge.
Does the result have sound?
Yes, if your motion video has sound: the original audio of the clip is kept, so music for a dance lines up with the moves. The model does not generate new sound; to add generated audio, use a model such as Seedance 2.5 in text-to-video or image-to-video.
Why was my motion control video rejected or failed?
Common causes: a missing image or video (both are required), a clip shorter than 3 seconds or longer than the orientation limit (10 s for image, 30 s for video), a file over the size limit, a character whose body is hidden or cropped, or content rules on real faces and copyrighted characters. Failed runs are not charged.

Open AI Image is an independent platform and is not affiliated with Kuaishou (Kling) or the other model providers named here. Upload only images and footage you have the right to use, and do not use motion control to impersonate a real person.

Related

  • Kling 3.0 model page
  • Kling 2.6 Motion Control model page
  • Video to video
  • Image to video
  • Reference to video
  • AI video extender
  • First & last frame to video
  • AI video price index
  • Pricing