How to make a clothing product video from photos with AI

Animate fabric motion, on-model scenes and lookbook edits from your garment photos. Morphic builds the motion around the real piece.

By The Morphic Team · Updated 2026-10-17

To make a clothing product video from photos, animate three clip types from your stills: a fabric-motion clip, an on-model scene and a lookbook edit. Each needs one clean still of the garment. Image-to-video builds subtle motion around the real piece, and you assemble the clips with captions and a music bed in Compose, then export the sizes each placement needs.

What you need

  • A clean still of the garment (a flat or hanger shot plus an on-model frame)
  • A one-line idea of the story each clip tells
  • A Morphic account

The three clip types for clothing

Clip typeWhat it showsExport size
Fabric motionThe material moving, a drape, a sway or a fold catching light9:16 reel, 4:5 feed
On-model sceneThe garment worn in a styled setting, generated around the real piece9:16 story, 4:5 feed
Lookbook editThe piece set in a short styled series9:16 reel, 1:1 product page

Fabric motion

What it shows: The material in motion, a drape, a sway or a fold catching light. Why it suits clothing: Weight and drape are what a flat photo loses. A short, restrained move shows how the fabric hangs while the print and cut stay exact.

On-model scene

What it shows: The garment worn in a styled setting, built as a generated scene around your real garment photo. Why it suits clothing: It shows fit and context without a studio shoot. This is a reference-driven generated scene, not a try-on of a shopper's body; the garment stays as photographed.

Lookbook edit

What it shows: The piece set in a short styled series, one look flowing to the next. Why it suits clothing: A lookbook frames the garment as part of a line and carries a drop or a collection across one reel.

Steps

Pick and save your stills

Choose the stills each clip will use: a clean garment shot and, for the on-model scene, a frame of it worn. Lift the garment off its background and save it as a reference so the fabric, cut and print stay accurate across every clip. Check that each still is sharp, since the motion inherits whatever the still gives it.

Animate each still with image-to-video

Animate each still into a short clip, describing the motion in plain words: a drape or sway for the fabric, a slow settle for the on-model scene, a gentle move between looks for the lookbook. Keep the motion subtle, because restraint keeps the garment accurate and dramatic movement warps the fabric and print. Check each clip for bending at the edges and regenerate the ones that distort.

Assemble with captions and music in Compose

Bring the clips onto the Compose timeline and order them into one cut. Add a caption track and generate a music bed for the clip. Trim each clip to its best seconds so the reel stays short and moves. Save the arrangement as a Workflow if you plan to make the same set for the next piece.

Export the sizes you need

Export the finished cut, then reframe the master to each aspect ratio: vertical 9:16 for reels and stories, 1:1 or 4:5 for the feed and the product page. Check that the garment reads clearly at small sizes and that captions sit inside the safe area for each platform.

Captions, sound and sizes

Captions carry a product video in a muted feed. Transcribe the cut to an SRT, translate it if you sell in more than one market, and burn styled captions onto the clip using a Latin-script font. For sound, generate a music bed for the clip rather than lifting a track you do not have rights to; the music is generated in the same workspace, so the reel stays clear for paid placement. For sizing, cut the reel vertical at 9:16 and reframe the master to 1:1 or 4:5 for the feed and the product page, so every placement matches instead of being rebuilt.

What to keep real

Image-to-video animates your real still, so the garment itself is never invented. Keep the fabric, the cut, the colour and the print exactly as photographed, and save the garment as a reference so it holds across every clip. The on-model clip is a generated scene built around the real garment photo, not a virtual try-on; it shows the piece in context, not on a shopper's body. Keep the motion subtle so the fabric and print stay accurate. The scene and movement around the piece can be generated; the garment stays the garment.

Start from the AI fashion model generator to build an on-model scene from your garment photo. For the still-image ad formats this method extends, see how to make a clothing brand ad. For more directions across categories, see product video ideas.

FAQs

How do I make a clothing video that keeps the garment accurate?
Animate a clean still of the real garment rather than generating it from a description. Save it as a reference, keep the motion subtle, and image-to-video builds the movement around the actual fabric, cut and print so the piece stays accurate.
Is the on-model clip a virtual try-on?
No. The on-model clip is a reference-driven generated scene built from your real garment photo, not a try-on of a shopper's body or a measurement tool. The garment stays as photographed; only the scene and movement around it are generated.
Which clips work best for clothing?
A fabric-motion clip that shows the material move, an on-model scene that shows the piece worn, and a lookbook edit that sets it in a styled series. Each starts from a clean still of the real garment.
What music can I use in a product reel for ads?
Use a generated music bed for the clip rather than a track you do not have rights to. Morphic generates the music in the same workspace, so the reel stays clear for paid placement without a licensing question.
What size should a clothing reel be?
Vertical 9:16 for reels, stories and short-form, and 1:1 or 4:5 for the feed and the product page. Make the cut once and reframe the master to each placement so the set matches.
Do I need video editing software?
No. Compose, the editor inside Morphic, assembles the clips, captions and music bed into one cut and exports every size. Nothing separate installs, and no timeline experience is assumed.