How Garu's AI sizing decides which size is yours.
Not a body scan and not a quiz. A short profile, matched against the specific garment on the page and the store's own size chart.
Sizing is a garment problem, not a body problem.
Every apparel brand already knows a shopper's measurements are not the hard part. The hard part is that a medium in one label is a small in another, that the same label sizes differently across a knit and a woven, and that a size chart describes the garment flat while the shopper wears it in three dimensions.
So Garu does not stop at the body. A shopper gives height, weight and body shape in a few seconds - no tape measure, no photograph required for a recommendation. That profile is then matched against the specific product on the page and the store's own size chart, which Garu reads rather than replaces.
Every recommendation carries a confidence score. Where the garment data is thin or the shopper sits between sizes, the widget says so instead of asserting a number, and that honesty is what keeps merchants comfortable putting it on the page. Roughly 42% of fashion returns are size-related, and a recommendation nobody trusts does not move that figure.
- Profile
- Height, weight and body shape, collected in seconds.
- Garment-specific
- The recommendation is for the item on the page, not a store-wide average.
- Your chart
- Mapped to the size chart you already publish, so the answer matches your labels.
- Confidence
- A score on every recommendation, including when the honest answer is "between sizes".