Program terms and spot request

The Q1 is delivered exclusively to companies joining AgnosPCB's PCBA supply program. It is not available as a standalone purchase or through any conventional sales channel. This exclusivity allows AgnosPCB to keep the unit cost at zero for participants: the collaboration itself is the exchange. Companies outside the program who wish to acquire a Q1 may do so through the standard commercial channels at the listed price.

The company commits to providing AgnosPCB with a minimum of 5 PCBA pairs per month, agreed prior to delivery and recorded in the application form. Each pair consists of two boards of identical design (no Golden Sample is required), simply two PCBAs from the same production run.

These boards are used exclusively to train and improve the inspection model under your specific production conditions: component types, board layouts, and defect patterns that are unique to your manufacturing environment. The more representative the boards, the sharper the model becomes for your use case.

Shipping is arranged by the participating company using the most cost-effective method available (no specific carrier is required). The handling of the boards once processed (return, destruction, or other arrangement) will be agreed individually between both parties prior to delivery.

The company maintains the monthly PCBA contribution for a minimum period of 12 months from the unit delivery date. This period ensures a sustained improvement cycle for the neural network. After 12 months, the collaboration may be renewed by mutual agreement under the same or updated terms, or terminated with 30-day advance notice. Early termination before the 12-month mark triggers the unit return clause (see condition 4).

If the company ceases to supply the committed PCBAs or breaches any term before the commitment period ends, the unit must be returned to AgnosPCB within 15 business days of notification. Return shipping costs are borne by the company. The unit must be returned in its original working condition; damage caused by misuse may result in additional charges. AgnosPCB will issue a formal written notification before initiating any return process, and both parties will have the opportunity to resolve the situation before proceeding.

Unit delivery is free of charge; the company assumes only the shipping costs to their premises. Ownership of the machine remains with AgnosPCB throughout the entire commitment period. This model allows electronics manufacturers to access industrial-grade AOI inspection capability without the capital investment that a conventional AOI system would require. If after the commitment period both parties agree to transfer ownership, terms can be negotiated separately.

The Q1 operates connected to the AgnosPCB neural network and requires a stable, permanent internet connection to perform inspections. Without connectivity, the unit enters standby mode and no inspections can be carried out. This cloud connection is also what keeps the Q1 always running the latest model version. No manual updates, no setup. AgnosPCB is not liable for inspection downtime caused by the client's internet outages. Companies requiring an offline solution may contact AgnosPCB to explore the AOI Offline product line.

The service processes 512×512 px image crops sent from the Q1 to the AgnosPCB cloud. All images are automatically deleted from the server every 24 hours (00:00 GMT) and file names do not identify the client or their products. The Q1 does not store full board images. Only the crops required for inference. Anonymized inspection metadata may be used to improve the neural network model. AgnosPCB does not share client data with third parties. Full details are available in the privacy policy.

Remote technical support is included throughout the commitment period at no additional cost. Failures attributable to normal use are covered by AgnosPCB; those caused by mishandling, accidental damage, liquid exposure, or external factors are the client's responsibility. AgnosPCB may access the unit remotely for diagnostics and software updates, always subject to prior agreement with the client. The Q1's neural network model is updated automatically via the cloud connection, ensuring every inspection runs on the most current version without any intervention required on the client's side.