About
FashionScout is a visual search and price intelligence platform for secondhand and archival fashion — built so that finding any garment across every marketplace takes seconds, not hours.
What it is
FashionScout lets you photograph any garment — from a runway look, an editorial, or your camera roll — and instantly identifies the piece using AI visual search. It then scans nine major secondhand and retail marketplaces simultaneously, surfacing the best price, the closest match, and arbitrage opportunities in one clean interface. No tabs, no manual searching, no wasted hours.
Our mission
We believe the secondhand market deserves the same intelligence tools as Wall Street. Our mission is to give fashion buyers, resellers, and curators a competitive edge through AI — making every search faster, every decision smarter, and every rare find more findable.
The technology
At the core of FashionScout is a specialised vision-language model trained on fashion — Marqo-FashionSigLIP — which generates 768-dimensional embeddings that capture garment silhouette, texture, colour, and construction with 57% better retrieval accuracy than general-purpose CLIP models. These embeddings power both Scout (visual search across marketplace listings) and Taste Engine (a personalised AI buyer that learns your aesthetic from every like and dislike, using max-similarity prototype scoring against a foundation of 98,000 catalogued runway looks). Garment descriptions are generated by Qwen3-VL-32B, a frontier vision-language model, enabling rich natural-language queries on top of visual similarity.
Scout
Upload a photo. AI identifies the garment, searches 9 marketplaces, and returns ranked results in under 10 seconds.
Taste Engine
An AI buyer that learns On9’s aesthetic from curator feedback — auto-scans marketplaces and surfaces matching archive pieces.
On9
Resale business tooling for sourcing, pricing, and tracking inventory across platforms from a single dashboard.
Built by