To make an AI dance video with your own photo in MoveFrame AI, you need two inputs: a clear portrait and an MP4 containing the dance movement. Upload the MP4, review the detected scenes, assign your portrait to the scenes you want to generate, then preview and download the result. A photo supplies appearance; the video supplies movement and shot order.
This guide covers the reference-video workflow in the AI motion transfer workspace. It does not require a text prompt, and it does not offer a photo-only choreography preset.
What do you need before starting?
| Input | Supported format | A practical first choice |
|---|---|---|
| Dance reference | Standard MP4, up to 100 MB | One person, stable framing, readable face |
| Your portrait | JPG, PNG, or WebP, up to 10 MB each | Clear face, similar angle and lighting |
| Audio | Audio already in the uploaded MP4 | Your own recording or music suitable for your use |
| Scene plan | Generate or keep each segment | Generate clear solo shots; keep difficult shots |
A short solo recording is easier to inspect than a full music video. For example, film a brief arm movement or gentle turn, with your face visible. This is a starting suggestion rather than a duration limit or a promise of a particular result.
If you have only a selfie, prepare a performance video first. A portrait cannot establish dance timing, camera motion, or a scene sequence. The reference footage checklist explains how to choose the second input.
Step 1: Upload your reference dance video
Sign in, choose an MP4, and start upload and scene analysis. Wait for the timeline before deciding what to generate. Inspect each original segment: confirm that the intended person is visible and check whether another performer enters the frame.
Prepare the intended framing before upload. For a vertical post, use footage already framed for that post rather than expecting the tool to create a different composition automatically.
Step 2: Add your photo and assign it to scenes
Add your portrait to the character library. Choose a scene and apply the portrait to it, or select several scenes and apply the same portrait to the selection. Apply to all generated scenes when the same person should appear throughout.
Different portraits can be used across separate segments. This does not provide separate identity controls for two people within one group shot. For a two-person concept, prepare separate solo segments or keep the group shot unchanged.
Choose Keep original for shots that should not be generated, such as titles, scenery, or a difficult overlapping dance. Review the generated-scene count and complete any missing portrait assignments before submitting.
Step 3: Generate, compare, and fix one segment at a time
Submit the completed scene plan. Preview individual results as they become available, then review the full video. Check the face through turns, hands passing in front of the face, and transitions between scenes.
Motion transfer can also change clothes, body details, or the background. Judge moving frames rather than only the thumbnail. If one segment needs another attempt, use its regeneration control. A failed task has a retry option. Persistent visual problems are a reason to simplify the reference rather than repeatedly submitting the same difficult shot; use the dance video troubleshooting guide to choose the next change.
Step 4: Check the soundtrack and download
The export combines segments in their original order and merges the source MP4’s audio. It does not automatically add a trending song, clone a voice, or guarantee singing lip-sync. Listen to the full preview and check the start and end before downloading the MP4.
Use the current workspace and pricing page for live availability and costs. A trial or price shown in an older tutorial may change.
Which creative tutorial should you follow next?
Use the Hotel Lobby AI video tutorial for a minimal studio performance, or the Chuttamalle AI video song tutorial for a personal dance edit. Each builds on the same MP4-and-portrait workflow, with different preparation choices.
Frequently asked questions
Can I make an AI dance video from a photo without a reference video?
This workflow requires an MP4. It does not animate a still image into preset choreography by itself.
Will my face stay identical in every frame?
Consistency is not guaranteed. Angle, blur, occlusion, and movement affect the result. Review each generated segment and keep or replace unsuitable footage.
Can I download the dance video with music?
The service merges the uploaded video’s audio. Prepare music you can use in the source file or select music separately in your editing or publishing app.
