Turn Any Clip Into a Depth-Map Control Video
Depth Anything V2 runs in this tab. Download a grayscale motion-control video, then attach it on a clip before you reshoot — the model follows camera and blocking instead of guessing from lyrics alone.
How to remake a viral MV shot
01
Convert locally
Drop the reference clip. Depth maps stay on this device until you choose to attach one.
02
Attach on the clip
In the clip editor, pin the depth video on a Seedance / Hailuo / Wan 3 row.
03
Reshoot
The new take follows the original blocking. Lyrics and character stills still drive identity.
FAQ
Can I turn a video into a depth map here?
Yes. This page runs Depth Anything V2 Small in your browser and exports a grayscale control video (near = bright).
Why not just upload the original clip as a reference?
A full-color reference invites the model to copy wardrobe and faces. A depth map keeps motion and camera, so you can recast the shot.
Which models use it?
Seedance (Fast / Mini / 2.5), MiniMax Hailuo 03, and Wan 3. Other clip models ignore the attachment.
Does this upload my footage?
The converter does not. Attaching the result for generation stores it like any other reference file so the model can fetch it.
How long can the clip be?
The first 15 seconds, 8–24 fps, up to 80MB. That matches Seedance’s clip window.
Is conversion free?
Yes. Spending credits only happens when you confirm a reshoot that uses the depth clip as a reference.
Reshoot with the depth clip attached
Open your song, drop the depth video on the clip, and generate. Identity still comes from your stills.
Open SunoMV