Complete the pairProtect the frame before it hits the bag.
The technology behind the Fit Desk
One fit measurement and local shopping signals turn hundreds of frames into one answer, a ranked store, or photo matches.
The Fit Desk does not ask a long lifestyle quiz and call the answer personalization. It compares your fit with published frame data, then uses your own local activity only to settle close calls.
Published width carries the ranking. Local activity can move a close match, but it cannot overpower a clearly better fit.
A daily starting pick from your top fit-ranked group, with the reason it rose.
My ShelfEight best matches first. The rest of the active catalog follows in useful batches.
Seen ItVisible frame shape turns into real matches you can shop and try on.
Inside the ranking engine
Every known-width frame is compared with your saved width. The strongest comfort range rises first. Your own activity can fine-tune the order, and the page tells you why each result moved.
The ideal frame is about 2 mm wider than your saved face width. Frames from 5 mm narrower through 9 mm wider stay inside the primary comfort band.
These are the actual local tie-break boosts. Their smaller weights are intentional.
Unknown-width and out-of-range frames are never erased. They stay available and are clearly labeled instead of being quietly hidden.
Results sort by score first and product handle second. That deterministic second step prevents equal-scoring frames from shuffling randomly between visits.
Pick a tool below. Each one uses the Fit Desk differently because choosing one place to start is not the same problem as rebuilding a catalog or matching a photograph.
One Frame ranks the live feed, takes your top eight, and gives you one daily starting pick. If it is wrong, “Not this one” moves through that top group without repeating a frame during the session.
My Shelf puts your eight best matches first, then keeps the rest of the active catalog available in eight-frame batches. Every card can explain why it ranked, while unknown widths remain visible as fit unconfirmed.
Seen It reads visible frame geometry from one uploaded photo. It does not identify the person or guess the brand. It returns catalog handles, then the storefront joins those handles to current product images, prices, fit context, and try-on links.
Inside Seen It
Seen It keeps the job narrow. It looks for the glasses, not the person wearing them. That lets it turn visual inspiration into live shopping options without identity recognition.
The browser reduces the longest edge to 1,000 px and steps JPEG quality down until the file is roughly 580 KB or less.
The matching worker compares the frame silhouette and proportions with the embedded catalog descriptors.
The storefront uses those handles to attach current images, product details, fit labels, Shop, and Try It On.
Private by design
The Fit Desk can remember what helps without building a hidden identity profile. Its useful memory stays visible, limited, and under your control.
Your saved width, views, try-ons, searches, and carts tune the experience on this device.
The compressed image is used for visible frame-shape matching, not identity, and is not retained.
Email and fit width leave the browser only when you explicitly request a My Shelf alert.
Start with one measurement
Save one face-width measurement. The Fit Desk uses it to sort real frames into stronger starting points across One Frame, My Shelf, and the catalog.
