Vinted bans edited and stock photos: what AI can still do

Vinted does not allow AI-edited listing photos. Its Catalog Rules say photos "have to represent the item as it is - no image editing is allowed", they must be sharp and clear, and you must take them yourself for the purpose of using them on Vinted. They "cannot be replaced with stock photos, photos from ads and watermarked images". The first photo has to show the whole item, and a listing can hold up to 20 photos.

This is from Vinted's Catalog Rules, read on 2026-10-02. The page says nothing about AI by name, but "no image editing is allowed" already covers generating, retouching or restyling a listing photo, so the rest of this post is about where Sume can help without touching the listing gallery.

What do Vinted's photo rules say?

The rules read as a short list of musts and must-nots.

What does that mean for a reseller using AI?

Do not use AI to clean up the background, fix lighting, remove a stain or restyle a flat lay in the photos you upload to Vinted. Take the photo again instead. A defect you hide is a defect the buyer finds, and the rule bans the edit regardless.

Even a light retouch is an edit by the plain reading of the page. If you are unsure whether a tool crosses the line, don't use it on listing photos.

Where can Sume still help a reseller?

Everything that is not the Vinted listing: a Reel announcing a drop, a story with your best pieces, a thumbnail for your own shop site. Sume's video API can turn your real item photo into a short 9:16 clip, and the Image API can lay it out as a clean 4:5 promo card. Both are asynchronous or can return a job envelope, so build your script to poll.

Do not reuse those promo assets as Vinted listing photos. Keep two folders: listing/ for untouched originals and promo/ for generated material.

import os, time, requests

H = {"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"}
job = requests.post(
    "https://api.sume.com/v1/videos",
    headers=H,
    json={
        "model": "seedance-2",
        "prompt": "Slow push-in on a folded wool sweater on a plain bed, soft daylight, the sweater unchanged",
        "frame_images": [{
            "type": "image_url",
            "image_url": {"url": "https://example.com/closet/sweater.jpg"},
            "frame_type": "first_frame",
        }],
        "aspect_ratio": "9:16",
        "duration": 5,
        "resolution": "720p",
    },
    timeout=60,
).json()
while True:
    time.sleep(20)
    st = requests.get(job["polling_url"], headers=H, timeout=60).json()
    if st["status"] in ("completed", "failed", "cancelled"):
        break
print(st["status"], st.get("unsigned_urls"))