YourSizer
An AI size recommendation that runs on the product page. The shopper builds a body model in under a minute; the engine matches it against the size chart of the specific product in front of them and returns one size — with a reason.
How it works
- 1Height and weight to startThe shopper enters two values. The remaining key body metrics are estimated from population data rather than asking anyone for a tape measure.
- 2Adjust the 3D body modelIn the size flow a 3D body model takes shape from the shopper’s measurements. Chest, waist and hip can be tuned directly and the model updates as they move.
- 3Match against this productThe recommendation is computed against the size chart that applies to this product — product override, then category, then brand — with the garment’s tolerance and the brand’s intended fit applied.
- 4One size, plus the caveatThe widget returns a size, and when the shopper sits between two it says so and names the measurement driving it. The size can be selected and added to the cart without leaving the page.
What the shopper never has to do
- Take a photo or upload a body scan to get a size
- Measure themselves with a tape
- Read a size chart or convert between EU / UK / US
- Rebuild their model on every visit — it is remembered on that store
What the merchant has to do
- Connect the store — app install, or one script tag
- Upload or import size charts, including from images
- Choose where the button appears on the product page
Under the hood
A calculator maps height and weight onto one chart. YourSizer resolves which chart applies, how tolerant that garment is, and how the brand intends it to fit.
Layered chart resolution
Product override → category → brand → system default. Published charts are immutable; editing one creates a new version, so an old recommendation can still be explained.
Tolerance resolver
How much a garment stretches or drapes resolves through a layered chain — system, category, brand, product — with a hard cap so no single layer can make a garment infinitely forgiving.
Fit-style multiplier
Fit intent is an explicit multiplier on tolerance: slim, regular, relaxed and oversize each scale it differently.
Adaptive measurement weighting
A dress is driven by bust and hip; trousers by waist and inseam. The engine weights the measurements that actually decide the fit of that category.
Children’s sizing
The European height-grid convention with a growth-percentile anchor — a child’s size is driven by height and age far more than by weight.
Charts from images
Charts can be ingested from an uploaded picture and turned into structured measurement rows for review, rather than retyped.
Virtual try-on
See it on yourself, not on a hanger
Alongside the size recommendation, YourSizer offers photo-based AI try-on: the shopper uploads one ordinary photo and gets an image of the product on themselves, on the same product page. Try-on is optional — a size never needs a photo.
Try-on is a separate entitlement from sizing. It can be enabled per store and per product — categories where an AI rendering would not be honest can be excluded — and there is a safety filter on generated imagery.
Try-on and the 3D model are not the same thing
The 3D body model belongs to the size flow: it is shaped by the shopper’s measurements, it is deterministic, and it is what the recommendation is based on.
The try-on image is generated from the shopper’s own photo to help them picture the product. It is a merchandising aid — it never overrides the size the engine computed.
What the dashboard tells you
Size distribution per SKU
Which sizes shoppers are actually being recommended for each product — the earliest signal that a production run is cut small.
Interaction funnel
Button clicks, time in the interface, how far shoppers get and where they drop, per store and per placement.
Assisted conversion
Orders where a recommendation or try-on was involved, deduplicated, with the signal that proved the link recorded next to the number.
Geo and device trends
Where your recommended sizes skew by region — useful before you commit a buy for a new market.
Return drivers
With ResReturn connected, size-related return reasons are joined back to the recommendation that preceded them.
Chart health
Products with a missing, stale or implausible chart are surfaced instead of silently falling back to a default.
Two ways in
On Shopify, ikas, Ticimax, T-Soft, IdeaSoft and Faprika the widget is installed through the platform integration: authenticate the store, sync the catalogue, place the button. No theme code for the common case. See platform coverage →
Built to sit on your store
Accessibility
WCAG 2.1 AA target: keyboard navigable, labelled controls, and no information conveyed by colour alone. On mobile it renders as an edge-to-edge bottom sheet, not a cramped modal.
It won’t fight your theme
Rendered inside an isolated root so your CSS cannot leak in and its styles cannot leak out, with stacking defences against third-party overlays.
Data protection
Body measurements are personal data and are treated as such: explicit retention, deletion on request, and per-store isolation.