Find it with a photo
Some products are hard to put into words. Visual search skips the words entirely: give it a picture and it finds the items in our catalogue that look most like it, across all 2,842 product photos we hold — usually in about a second.
Give us a picture
Drag a photo onto the box, click to browse, or paste a screenshot straight from the email your client just sent. Phone photos are fine — so are HEIC files off an iPhone.
We read the picture, not the words
An open-source image model turns your photo into a visual fingerprint and compares it with the same fingerprint taken from every product photo in our catalogue. No keywords, no category, no SKU.
Closest matches, best first
You get the 24 nearest-looking products ranked by similarity, with live pricing and stock — ready to add to a kit, a flyer or a virtual.
When it earns its keep
“Can you get something like this?” — find our closest equivalent before you reply.
Barrel bag? Dopp kit? Skip the vocabulary and search the shape.
Pinterest, a competitor's PDF, a trade-show photo — paste it and see what we carry.
Building a themed gift set and need pieces that sit together visually.
Getting a better match
- One product per photo. A shelf of six things gives the model six answers to average.
- Fill the frame. Crop to the product before you upload — corners full of desk and hands pull the match around.
- Plain backgrounds win. A catalogue-style shot beats a lifestyle photo almost every time.
- Shape beats colour. Matches are found on form, texture and proportion, so expect the right item in a different colour — most of ours come in several.
Where your photo goes
Nowhere. The image is processed on our own server by a self-hosted open-source model, turned into a list of numbers, and dropped — it is never written to disk, never sent to Google, OpenAI or anyone else, and never added to our catalogue.
Found it? Put it straight to work — add your client’s logo, drop it in a gift set, or build a flyer around it. More self-service in the Pre-Sales Tools.