VID · Motion, explained
Fashion video, generated — the governed way.
Motion converts better than stills on every surface it touches — and generation makes it affordable for the whole catalog, not just hero SKUs. This guide covers how our motion lane works, the craft of directing movement, and why you never pick an AI model here.
Step by step
Three decisions. None of them is a model family.
01 · Start from a governed render
Your source is an approved production image — one that already passed the consent gate and the art-director review. Motion inherits the clearance instead of reopening it.
02 · Pick a motion template
Turnaround, walk-in, breeze, or orbit — motion language a fashion buyer recognizes. Add a line of direction only if you want to; the template carries the craft.
03 · Run — quoted, with audio
The exact credit cost sits on the Run button. Audio generates natively in the same pass. Failed runs refund themselves.
Coverage
What the templates ship
Turnaround
The full 360° — the PDP motion staple. Garment drape through the turn is the whole point.
Walk-in
The model walks to camera — campaign energy for ads and Reels covers.
Breeze
Fabric moves, model nearly still — the editorial calm that sells texture.
Orbit
The camera circles a held pose — lookbook luxury with the garment in frame throughout.
Duration guidance: 6 s for product pages · 8–10 s for ads and Reels. Ratios: 9:16 vertical, 16:9 for product pages and site heroes. 1080p output.
Direction
Direct the motion, not the scene.
On an image-to-video run the first frame already fixes the subject, outfit, and set. A good direction line answers up to four questions — every one you answer takes a decision away from the machine:
Subject motion
What moves, and how — “she turns slowly”, “the coat sways”.
Camera
Name the move or the engine picks one — “static”, “slow zoom”, “orbit”.
Pacing & mood
“Slow, calm movements” vs “energetic” — pacing makes a clip feel intentional.
Constraints
What must not change — “same pose”, “keep the garment fully in frame”.
One clear motion per clip — a 6-second video cannot fit five actions. And if the result should differ in scene, garment, or model: that is a production change, made upstream where the rights rail governs it, not a motion prompt.
The picker you won't find
You shouldn't need to become a model sommelier.
Self-serve tools hand you six engine families and a capability matrix — audio on some models, end-frames on others, ratios varying per engine — and make the choice your problem. We treat engine selection as an engineering decision the platform already made:
Native audio, one pass
The primary engine generates synchronized audio with the motion — no toggle, no second step.
Automatic fallback
If a generation fails, a proven fallback lane covers it — and every result reports which engine delivered.
One price, quoted
No per-model credit tables. The cost is on the button; failures refund themselves.
FAQ
Common questions
How much does an AI fashion video cost here?
A Motion Studio clip is 350 credits, shown on the Run button before you generate. Credits debit before generation and are refunded automatically if a run fails — there is no variable-cost surprise and no debt.
Which AI video model should I choose?
You don't choose one. The platform runs a deterministic engine choice — LTX 2.5 for native synchronized audio, with an automatic fallback lane if a generation fails — and every result reports which engine produced it. Picking between model families is our job, not yours.
Can the videos have sound?
Yes — audio is generated natively in the same pass on the primary engine, synchronized to the motion. No separate audio step, no toggle to remember.
Do I own the videos I generate?
Deliverables from governed productions carry their clearance with them: the XMP rights passport (consent token, talent ID, consent scope, royalty event) is embedded in the source render, talent royalties are recorded at approval, and the CO₂e certificate is publicly verifiable. That is a stronger position than a terms-page permission — the clearance is part of the asset.
What should a motion prompt say?
Direct the motion, not the scene: what moves, how the camera behaves, the pacing, and what must not change. The first frame already fixes the subject, outfit, and set — most templates need no prompt at all.
How long should a clip be?
6 seconds is the product-page sweet spot; 8–10 for ads and Reels where one motion needs a beat to land. Match length to the destination — longer costs more and rarely converts better on a listing.
Motion for the whole catalog — with the rights inside.
Start with one garment and get a rights-scoped production plan; motion rides on the renders it produces.