Bring a character image, keep the preview as the performance to follow, and review the first take before you extend it. The check is identity plus timing, on a single take.
Transfer this performance
Add a character image and follow the move in the preview.
Motion preview
The clip on this page is the performance to follow
Transfer brief
Add a character image, follow the preview’s performance, then check identity and timing before a longer export. Say which part must stay fixed, such as the face or the clothing color, and leave a second routine out of the brief.
01Choose one image where the face, torso, and limbs are easy to see. Front or three-quarter views usually transfer more cleanly than a distant crowd. Match the reference’s body orientation when you can. A profile photo against a front-facing performance is a common reason the first take twists. Note the outfit you want to keep. This step sets who performs. It does not outline a multi-scene film.
02This AI Motion Transfer template uses the clip on this page as the performance to follow. Write one line for camera distance, outfit, and energy, and say which part should stay fixed, such as the face or the clothing color. Leave celebrity names and brand logos out of the brief.
03Play the short result and watch the face, hands, and whether the timing still matches the preview. If the person drifts, simplify the image or the action and generate again. Change one instruction on the retry, such as camera distance or how much of the body stays in frame, instead of rewriting the whole brief. Open a model workspace when you need orientation or original audio. Credit use follows the allowance shown for your account.
This page stays on one transferred performance. Open a motion-control workspace for orientation and audio, or a dance template when the job is a routine rather than a general move.
What lands on your character, what you can change on this page, which controls sit in the model workspace, and when a dance template is the narrower job.
This page places the preview’s performance onto a character you upload. You add an image and a short brief for camera, outfit, and what must stay recognizable. The result to judge is one take: your subject, moving with the reference timing. It transfers a performance. It does not rebuild the reference shot as a new story, and it does not promise that every finger matches on the first try. Watch the frame where the reference changes direction. If an arm merges there, shorten the action before you extend. Accept or revise after that first watch. Name the body part that must stay recognizable, then look only at that part on the first take. If you said the face, ignore a tempting camera move until the face holds. A cluttered background makes the silhouette harder to separate from the set, so a plain wall is the fairer test.
You need a character image with a readable face and body, and a performance idea that fits the preview. Solo, clearly lit movement transfers more cleanly than a crowded group. A front or three-quarter view beats a profile when the reference faces the camera. Confirm the file types shown in the recreate panel before you upload. The clip on this page is the motion source. Your image is the performer. Only upload a person when you have the right to depict them, and only follow a reference move you have the right to use. If the reference is a long routine, pick the clearest few seconds first so the first take is easy to judge. If your still is seated and the preview is a walk, the first take has to invent legs that were not in the frame. Start from a still that already shows the limbs the performance uses. Write the clothing color down and check it when the motion stops.
On this page you can replace the character image and rewrite the brief. The preview stays the performance to follow. For model settings, open the orientation, audio, and duration controls. When that panel shows them, Character Orientation is Image or Video, and Keep Original Audio is a switch. The same workspace lists 480p and 720p, aspect ratios including 16:9, 9:16, 1:1, 4:3, 3:4, and 21:9, and a duration range of 5 to 20 seconds. Confirm those values on the live panel. This template does not redefine what Image versus Video orientation does beyond those labels.
Stay here when you want a general performance moved onto your character. Open the dance video template when the reference is a routine and you want the brief, framing, and review focused on footwork. A dance page still starts from a dancer image and a preview. It is a narrower job. Do not force lyrics, chapters, or a music-video storyboard into this transfer brief. Footwork-heavy references belong on the dance template, where the review is the feet. A gesture, a sit, or a turn can stay here. Do not add lyrics to make a general move behave like a music video. That adds a story the preview is not showing.
This URL sets the character and the brief. Duration, resolution, ratio, orientation, and original audio are selected where the model runs, and only when those controls are enabled. A 5 to 20 second range, 480p and 720p, and the listed aspect ratios are workspace values to confirm on the live panel, not promises printed on this template. Keep Original Audio, when the switch is on, plays the reference sound. It does not license that soundtrack. Read the per-second cost in the create panel before you pick a longer duration. Buzzy shows a start-for-free entry, and generation still follows the credits and plan limits in your account.
Transfer drifts when the body in the image and the body in the reference do not line up. Faces soften, an arm merges into the torso, or the timing slips. A short first take shows whether identity and the move survived. If the person is right and the action is muddy, change only the motion line. If the face no longer matches the upload, simplify the image before a longer generation. This model can be priced by the second in the create panel. Read the cost shown for your account before you extend the clip. A second generation should change one instruction, such as camera distance or how much of the body stays in frame, instead of rewriting the whole brief. Line the bodies up before you blame the brief. A profile still against a front-facing performance twists the shoulders even when the sentence is clear. Swap to a closer angle or shorten the action, and leave the other sentences alone so you can see what changed.
Follow Buzzy’s current terms for generated media. A transferred clip is still a generated performance under those terms. The AI Motion Transfer template does not give permission to copy a choreographer’s routine, a celebrity likeness, or a branded character you do not represent. Review likeness rights for the person in the upload and rights for the reference move before you publish. Keep Original Audio, when you leave it on in the workspace, does not grant a license to the source soundtrack. This page does not promise unlimited exports or that every resolution is included. Confirm usage rules in your account if the clip will run in an ad or on a channel.
Open the AI Motion Transfer preview, add a character image, and confirm the face still matches before a longer performance.
