A character illustration already tells you what someone looks like. A performance tells you how they move: the pause before a wave, the timing of a step, or the energy of a turn. Combining those two ingredients is the creative idea behind motion transfer.

Motion Control AI lets you pair a character image with a reference video to generate an animated clip. The image supplies the visual starting point, while the video guides the performance. An optional prompt can refine the appearance around that movement.

Choose the performance before writing the prompt

Begin by deciding what your character should communicate. A welcoming mascot might need a calm wave. A music teaser might need a short dance phrase. A concept scene might call for someone taking a step forward and looking around.

Find or record a reference that expresses that intention clearly. For a first experiment, choose one visible performer, a steady camera, and a short action with a clear beginning and end. A complicated montage gives you more moving parts to evaluate and makes a weak result harder to diagnose.

Match the character image to the movement

Think about which body parts the action requires. A close-up portrait provides little visual information for a full-body dance. For a stepping or turning sequence, a clear image showing the relevant limbs is a more sensible starting point.

Compare the image and reference side by side. Are the poses broadly compatible? Can you see the hands when they matter? Does the character have room to move without immediately leaving the composition? These questions help you choose inputs before spending time on revisions.

Build your first clip

Upload the character image and motion reference together. The current homepage accepts JPG or PNG character images and reference clips in MP4, MOV, or MKV. It displays reference-length limits according to the selected orientation, with support for up to 30 seconds in the appropriate configuration.

Review the model, orientation, quality, and generation cost shown in the interface. Sign in for the credit-based workflow and check your available trial allowance. Treat a short first generation as a creative test, then review the complete output before deciding whether to make a longer version.

A practical mascot example

Imagine a small coffee brand with an illustrated barista character. The intended clip is a greeting for the opening of a social video. Record a simple reference: stand still, lift one hand, wave once, and settle back into a neutral pose.

For the optional appearance prompt, try an instruction such as:

“Keep the illustrated barista recognizable, with the same apron and warm color palette. Simple café background, gentle lighting, clear silhouette.”

The reference already supplies the action. Use the prompt to clarify the look you want, then inspect the animation for hand distortion, changes in clothing, and unwanted background movement. This is a proposed workflow, not a claim that every source pair will produce the same result.

Make revisions you can learn from

If the movement looks confused, first inspect the reference. Fast turns, crossed limbs, or a performer leaving the frame may make the action harder to interpret. If the character changes too much, try a clearer source image and a simpler performance.

Change one input at a time. Keep a note of the source pair and settings for any clip worth reusing. That gives you a practical reference library for future posts instead of a folder of unrelated experiments.

Watch the result once for the overall performance and again for details. Check whether the greeting feels friendly, whether the character stays recognizable, and whether the final pose gives you a clean place to cut.

Start with one expressive action

A short gesture can be enough to give a still character personality. Bring a clear image and a focused reference performance to Motion Control AI, then develop the clip through careful review. The most useful first result is one you can understand, refine, and repeat.