Keep the world in your photo
Choose the person, clothes and setting in one image. A clear, unobstructed body gives the animation more information to work with.
AI Motion Transfer
Your photo. Their moves. Animate the person in your image with movement from a video, using your photo as the guide for their look and setting.
- 1 photo + 1 motion video
- 3–15 seconds
- Your photo sets the scene
What is AI motion transfer?
AI motion transfer uses a video's body movement to animate a character in a still image. The image supplies the person's appearance and visual setting; the video supplies the action and timing. This reference-driven workflow is also called motion control.
Borrow the performance
Use a dance, a wave or another visible body movement. Pick one performer in a continuous shot, with the body parts needed for the action in frame.
It generates new frames. Your photo guides the face, outfit and background, but fine details, framing and lighting can change. It is not a pixel-locked background edit.
How to transfer motion to your photo
- 01
Add a motion video
Upload a 3–15 second clip or choose a supported motion template. Trim to the exact action you want before uploading; longer uploads use the first 15 seconds.
- 02
Add your character photo
Upload one clear photo with the outfit and background you want to guide the result. Match the photo's framing to the action: full body for footwork, upper body for gestures.
- 03
Generate, review and download
Check the credit total, generate, then review the face, hands, feet and background in your result. Your video appears in History for preview and download.
Motion transfer or character swap?
Choose by what you want to keep. Motion transfer animates your photo; character swap recasts a video.
| Workflow | Person and appearance | Setting | Movement |
|---|---|---|---|
| Motion transfer | Guided by your photo | Guided by your photo | Taken from the motion video |
| Character swap | Replace selected video characters using photos | Keep the source video's scene | Keep the source video's performance and cuts |
| Text-to-video animation | Described in a prompt or guided by an image | Described in a prompt or guided by an image | Generated from a written description |
Choose inputs that work together
For your photo
- One person, a visible face and clear body proportions.
- Show hands and feet when the action needs them; leave room around the body.
- Use the outfit and setting you want. Avoid busy foreground objects and heavy filters.
For the motion video
- One clearly visible performer, in a continuous shot.
- Keep the head and moving limbs in frame. Avoid fast cuts, crowds and long occlusions.
- Choose a pose and camera angle similar to your photo. Start with simple, readable movement.
More than a dance generator
Dance practice and remixes
Apply a short dance phrase to a new character without writing choreography into a prompt.
Gestures and reactions
Try a wave, a pose change or an upper-body reaction with a clearly framed reference.
Character animation
Explore a human-shaped character performing a recorded action. Stylized characters need testing; matching anatomy helps.
Motion transfer questions
Is motion transfer the same as motion control?
Here, motion transfer describes moving an action from a video onto an image character. Motion control is a broader term that can also mean controlling a camera or a motion path. This tool uses a reference video for the person's movement; it is not a camera-path editor.
Will my photo's background stay exactly the same?
The photo, not the motion video, guides the background and the person's appearance. However, AI generates new frames, so textures, objects, light and framing may shift. For a background that must remain pixel-for-pixel identical, you need a separate compositing workflow.
How long can the video be?
This page uses the existing 3–15 second motion workflow. Output follows the processed reference length; uploads longer than 15 seconds use the first 15 seconds. Trim the source yourself if you want a later section. There is no long-video stitching on this page.
How much does motion transfer cost?
The current rate is 10 credits per output second, rounded up to whole seconds: a 5-second clip is 50 credits and a 15-second clip is 150 credits. The generator displays the total before submission. Use the site's existing credit balance and plans.
Do I need a prompt, a full-body photo or a specific resolution?
No prompt is required. Use a full-body photo for footwork and an upper-body photo for hand or arm gestures. Images must be JPG, PNG or WebP, at most 10 MB and at least 300 pixels on each side. This page uses the existing 720p output preset, with no separate resolution selector.
Does it copy the music and lip sync?
The workflow requests the reference video's original audio. The public how-to preview is muted. Body-motion transfer does not guarantee precise singing, speech or lip synchronization. Only use audio and footage you have permission to use.
Can I animate several people or pets?
Start with one human or human-shaped character and one clearly visible performer. This page does not provide separate role mapping for multiple people, and animal anatomy can behave unpredictably. Use Character Swap for separately assigned video characters or browse the dedicated pet pages for examples.
What happens if generation fails?
The task uses the same status, History and credit-handling flow as the homepage. Confirmed generation failures return the charged credits through that flow. An output that finishes but has visual artifacts is not automatically classified as a failed task; review it before making another paid attempt.