Snap a product as it arrives. Know it in seconds.
Shelfsight turns your recognition API into a clean, responsive screen your shop staff use on a phone or a desktop: upload a photo, read the category and details, move on.
- 2.6 s
- avg. recognition
- 94%
- model accuracy
- 7
- categories
What staff get
Everything between the photo and the shelf.
Snap or upload
Live camera capture on mobile, drag-and-drop on desktop, with format and size checks before anything is sent.
Result in seconds
Product, category, confidence and details returned by your AI, laid out to be read at a glance.
One-tap correction
When the AI is unsure, staff fix the category or product in one tap. The correction goes back to your model.
Category overview
A browsable view of every category with the products recognized in each one, with filters.
Searchable history
Every arrival with date, photo, category and operator. Find any past product in seconds.
Plugged into your REST API
Upload, results and categories all go through your backend, with clean loading and error states.
How it works
Three steps at the delivery point.
Photograph the product
At the delivery point, straight from the phone camera.
The AI recognizes it
Your model returns the product, category and confidence.
Confirm and shelve
Staff confirm or correct, the back office is updated live.
POST /api/v1/recognitions → 200 OK · 2.4 s
2 interfaces, one application
Choose your role to explore the prototype.
