Motion designers
Explore a new subject performing an existing filmed action.
Film & CG workflow · depth-guided-motion-transfer
Video Motion Transfer AI
SEELE provides video motion transfer through its depth-guided workflow: one source video supplies performance, depth, camera, timing, composition, and audio guidance, while one attached image or text description defines the replacement subject.
Start in Film & CGNo prompt transfer is claimed from this static page.
Use cases
Built for a concrete production decision
Each use case stays inside the mapped Workspace capability rather than promising an unsupported all-in-one pipeline.
Creative technologists
Prototype depth-guided replacement with an explicit input contract.
CG teams
Evaluate whether source motion and camera survive subject replacement.
Workflow
From input to a reviewable artifact
The visible steps below match the HowTo schema and the supported Film & CG guide.
Supply one video
Attach the sole source performance and camera reference.
Define one subject
Use one attached image, or text only when no image is attached.
Derive depth
Convert the source to depth-spatial guidance while retaining timing and audio.
Run and inspect
Generate from the subject reference plus depth artifact, then review failure points.
Prompt
Start with constraints, not adjectives
Use the canonical brief as a working draft. Add only facts, references, and delivery requirements you can verify.
CANONICAL BRIEF
Transfer the performance from my one attached reference video to the replacement character in my attached image. Preserve timing, camera movement, composition, and audio; use the derived depth video as the only motion video supplied to generation.
Transfer the performance from my one attached reference video to the replacement character in my attached image. Preserve timing, camera movement, composition, and audio; use the derived depth video as the only motion video supplied to generation.
Variant 01
Humanoid performance to stylized character
Humanoid performance to stylized character
Variant 02
Creature replacement from a text description
Creature replacement from a text description
Variant 03
Prop or object replacement with preserved camera
Prop or object replacement with preserved camera
Outputs
Know what the workflow can hand back
Output names describe the evidenced artifact or delivery preference; exact file formats must be checked in Workspace.
Transferred-motion video
Replacement-subject result guided by the source performance.
Depth guidance file
Derived depth.mp4 passed into generation.
QA observations
Notes on pose, silhouette, occlusion, timing, camera, and audio.
FAQ
Questions before you open Workspace
What does this video motion transfer AI do?
SEELE provides video motion transfer through its depth-guided workflow: one source video supplies performance, depth, camera, timing, composition, and audio guidance, while one attached image or text description defines the replacement subject.
How quickly can I start?
Open the Film & CG workspace, provide the required brief and attachments, then run the mapped capability. Generation time varies with the request and service conditions.
Can I edit the result?
You can revise the brief and rerun the workflow. Any exact editing, export, or downstream integration must be verified in the resulting Workspace flow rather than assumed from this landing page.
Do I need source files?
Yes. Exactly one reference video is required, plus one replacement-subject image or a non-empty text description when no image is attached.
What should I check before production use?
Review factual accuracy, rights, identity, continuity, motion, framing, audio, and technical delivery. Generated media can require manual polish.
Limitations
What still needs judgment
These constraints are part of the product answer, not fine print.
Not generic mocap exportThis workflow produces guided video, not a promised skeletal animation file.
One-video ruleThe original video is converted to depth and is not also sent as a second generation video.
Manual QA is essentialAnatomy, contact, identity, and occlusion may need another pass.
Film & CG workspace
Turn this brief into the next reviewable artifact.
Open the mapped depth-guided-motion-transfer workflow, add any required source material, and keep the limitations visible while you iterate.
Open Film & CG Workspace