Vestiaire Collective has standardised the main photo of every new listing using Photoroom's AI editing system after a test showed a 5% increase in liquidity.
The luxury resale marketplace split new listings into two random groups. One was processed by Photoroom and the other by an in-house photo clipping tool developed several years ago. Listings edited by Photoroom sold at a higher rate, while product pages drew 3% to 7% more unique views. Bags and shoes showed the strongest gains.
The change applies to the main photo on each new listing before it goes live. Sellers still upload images as usual. The system then removes the background, adds a shadow, centres the item and adjusts lighting by product category.
For bags and shoes, it applies a deeper shadow to make the shape more visible. Clothing photographed flat receives little or no shadow.
Authenticity concerns
The test addressed a central issue for resale platforms, which rely on photography not just for merchandising but also for trust and verification. Vestiaire Collective reviewed several generative AI tools before selecting Photoroom and found that other systems altered product details.
That matters because the marketplace authenticates some items from photographs alone, checking labels, stitching and how light falls on materials. Buyers can also return goods that prove to be counterfeit or have undisclosed faults, making image accuracy both a commercial and operational concern.
"We tested different GenAI solutions, and each time they changed some details. Then it was no longer possible to authenticate from the pictures or to make sure buyers could trust them. Photoroom maintained product fidelity, and that's clearly the reason we chose it," said Thomas Hezard, Chief Product Officer at Vestiaire Collective.
Digital authentication was unaffected during the trial. The same images were also used in performance marketing adverts, where Vestiaire Collective recorded a 2% improvement in efficiency.
Scale of rollout
The marketplace operates in 70 countries and adds about 40,000 listings a day, ranging from private sellers taking pictures at home to semi-professional sellers using AI tools to improve their images. The test moved from the initial decision to results in about eight weeks.
Photoroom said the deployment is part of its enterprise image-processing service, which combines its own models with third-party open models. The system used for Vestiaire Collective includes FLUX.2 [klein] from Black Forest Labs and DINOv2 from Meta.
The service runs on NVIDIA H100 and H200 graphics processors, with capacity from Nebius and inference optimisation using NVIDIA TensorRT and NVIDIA Triton Inference Server. Some models receive only a few requests each hour, so they shut down when idle and restart when traffic returns.
Photoroom also said it worked with NVIDIA on Dynamo Snapshot, cutting the time needed to bring a new GPU server online from about three and a half minutes to about a minute. This helps absorb sudden spikes in listing volumes.
Wider market signals
Photoroom linked the Vestiaire Collective results to broader concerns about image consistency and trust in online marketplaces. In its own benchmark of 4,250 AI-generated images across 850 real products, the company said the best image-editing models passed a full product fidelity check only 29% of the time.
It also cited survey data showing that 63% of UK consumers say inconsistent product images or branding make a marketplace less trustworthy. For resale platforms, where condition and authenticity affect pricing and purchase decisions, even small visual changes can have outsized effects.
Hezard said Vestiaire Collective had compared its presentation with that of primary luxury retailers before making the change. "We looked at how first-hand players create emotion to convince a buyer to purchase a €1,000-plus item. Then we looked at our photos, and the gap was clear," he said.
Photoroom framed the project as evidence that marketplaces may no longer need to build and maintain specialised image tools internally. "Nobody returns a photo. They return the product. Vestiaire Collective tried the tools that change products and picked the one that doesn't, and more of its items now sell. Marketplaces have spent years building their own tools to get this right, and we want them to be able to buy it instead," said Matt Rouif, Co-founder and CEO of Photoroom.