Can’t find yours? Write to us and we will reply within two business days.
Do I need to know how to code?
Not to label, check the data, train or evaluate: it is all done in the browser. Connecting the model to your line takes one API call or the edge runtime, with examples in five languages.
How many images do I need?
It depends on the part and the defect. The data diagnostics warn when a class has few images, and the evaluation tells you whether there are enough test images to measure reliably.
Where is my data?
In your own instance, hosted in the EU, with daily encrypted backups. Training uses cloud GPUs. If the images cannot leave your company, Ocelo Local runs on your servers, offline too.
Can I take my data and my models with me?
Yes. Images and labels export as COCO, YOLO or PNG masks, and models as ONNX with their manifest.
Does it connect to my PLC or my MES?
Through the API or the edge runtime, with a standard shop-floor gateway: a line PC, Node-RED, or OPC UA or MQTT middleware. Ocelo does not include native PLC or MES connectors.
What doesn’t Ocelo do?
It does not read text or codes (OCR), it does not work with 3D point clouds, and it does not track objects across video frames. It works with 2D images: the camera or the line PC sends each image and gets the decision back.
What if I need more than Startup includes?
Write to us and we will extend your instance: more users, storage, GPU hours or deployments.
Where is it used today?
Ocelo is in production in human and veterinary medicine. The same core serves industrial visual inspection.
Do you have a visual inspection case?
Tell us what the part is, what defect you are looking for and how you take the photos today: we will reply whether Ocelo fits and how to start.