A row of drinks sells out at four on a Friday. Nobody notices until close. Every customer in between reached for a door, did not find what they came for, and bought nothing — or bought it somewhere else. That gap is the whole reason this exists.
There is a version of this product with POS integration, SKU catalogues, planograms and a dashboard to interpret. It would be more impressive to demo and much worse to own. Shops have high staff turnover and no time; anything that needs training or maintenance quietly stops being used.
So CoolerScout has exactly one setup step — show it a full cooler once — and its output is a sentence naming a cooler and a shelf. No integration, no catalogue to keep in sync, nothing to learn.
The camera does not move, so every shelf is a fixed rectangle in the frame. Each one is compared against a photograph of that same shelf when it was full — which cancels the shelf rails, price tags and glare that are in every picture whether or not there is product behind them. It needs no model and no training data, and it works the same on cans, cartons and jars.
An empty shelf is unmistakable. Telling nearly out from fine is genuinely harder, and we would rather warn you slightly early than let a row sell out in silence. A camera that gets knocked out of alignment needs recalibrating, and the app will say so rather than quietly reporting nonsense.
Built in California. Shipping across the US. support@coolerscout.com