Straight answers to what manufacturers ask us most: how much training data Aletis needs, what cameras work, the defects it catches, how long setup takes, and what it costs.
No. Aletis trains on good parts only. You never collect, photograph, or label defect examples. It learns what a correct part looks like and flags anything that deviates, including defects it has never seen before.
About 25 good parts. That is the entire training set. No thousands of images, no labeled defect library.
No technical expertise required. We promise. Your IT or operations manager sets it up. There is no data-science project and no model tuning.
Yes. Aletis works with your existing USB, GigE, or IP cameras. If you would rather not source hardware, the standardized Aletis kit ships ready out of the box with a matched camera, lens, and lighting.
Aletis reliably catches defects from 2 mm. Smaller defects are possible depending on your camera and lighting.
Anything a trained human eye is used to checking: surface defects, scratches, missing parts, and print or label errors, among others. If a person can spot it by looking, Aletis can be trained to catch it.
Under 30 minutes, including training, done by your own IT or operations manager.
Seconds. Each part gets a pass or fail verdict fast enough to keep up with your line.
Yes. Aletis adapts to your existing line and setup rather than forcing you to rebuild around it.
Traditional machine vision is supervised: it needs thousands of labeled defect images and a retrain for every new defect type. Aletis learns from about 25 good parts instead, needs zero defect samples, and catches defects it was never shown.
See the full comparison of anomaly detection vs machine vision →Pricing is flexible and matched to your line. Choose a monthly subscription with equipment leasing, or a fixed one-time price. Contact us for a quote tailored to your setup.