Boutique dress try-on video from one garment photo on Sume

A boutique can make a try-on video by calling the catalog Format sume-virtual-try-on with a garment photo and a person photo. Sume describes it as making a vertical video of a person wearing a supplied garment, true to its fit and construction, for try-ons, outfit previews and fitting videos for apparel.

It is a preview tool for a store that cannot shoot every new arrival. It is not a size guide, and the video should never be the only thing a customer sees about fit.

Pick the right Format

Sume has two nearby Formats. sume-virtual-try-on is for a person wearing a supplied garment in a vertical video. sume-virtual-fitting focuses on garment fit, silhouette and movement on a supplied person, for ecommerce fitting previews and product-page videos. Not for: a still lookbook or a garment-only product shot, per the Format description for try-on.

A boutique that wants a social preview of a new dress picks try-on. A shop that wants a product-page clip about how the cut moves picks fitting. Read both with GET /v1/formats/sume/{slug} to see the io profile before you decide.

The inputs and the consent rule

Attach up to 30 images on a Format run. For a try-on, that is the garment photo on a flat surface or hanger, and a photo of the person who will appear in the video. Use a staff member or model who has agreed in writing. Do not put a customer or an influencer into a video without permission.

Send a clear garment photo: the whole item, even light, front view. A photo with a busy background or a folded sleeve gives the Format less to be true to.

curl -sS -X POST "https://api.sume.com/v1/formats/sume/sume-virtual-try-on/runs" \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: dress-0142-tryon-v1" \
  -d '{
    "instruction": "Vertical 9:16. Show the dress on the model, a slow turn, then a close look at the neckline.",
    "attachments": [
      { "type": "input_image", "image_url": "https://example.com/dress-flat.jpg" },
      { "type": "input_image", "image_url": "https://example.com/model.jpg" }
    ],
    "generation_spend_cap_usd": 10
  }'

Check the output like a returns department

Pause on the hem, the sleeves, any print and any hardware. A generated try-on can drift at those details. If the video shows a length or fit the dress does not have, do not post it. Customers who buy against a video that misrepresents the item send it back.

Keep real photos and a size chart next to the video on the listing.

Many dresses at once

For a whole new collection, the same Format can run in a bulk queue of up to 100 items, with your concurrency setting. Set a generation_spend_cap_usd per run and review every output before it goes live.