AI can now run on devices without setup. From sensors to PCB inspection, these tools help electronics teams build faster and test instantly.
Many teams struggle to deploy AI applications quickly while ensuring performance. SECO’s latest release addresses this by providing ready-to-use solutions validated within its hardware ecosystem, so client teams can focus on product work rather than building AI from scratch. The update also adds a “Deploy” button, offering access to a 90-day trial of the Clea instance. All applications are documented in the SECO Developer Center, with deployment instructions, input/output specifications, and step-by-step testing guides for adoption.
The new applications cover a range of use cases. Llama Local Chatbot with Intel OpenVINO provides a locally-run conversational AI for edge devices, solving on-device chat without cloud dependency. Sensor Forecast Smart Building (NeuralProphet) predicts sensor patterns in buildings, helping teams manage energy use and plan maintenance. Face ID + Edge Impulse provides facial recognition for developers using Edge Impulse, simplifying edge ML deployment.
For visual inspection, Scene Text Detection High Accuracy & Light Model offers two OCR options: a model for precise tasks and a model optimized for speed on limited devices. PCB Defect Detection identifies flaws in printed circuit boards, streamlining quality control in manufacturing. YAMNet Training Experience provides a framework for training audio event classification, allowing users to recognize custom sounds for their applications.
These solutions help organizations integrate AI efficiently, reducing development time while maintaining performance across industrial and edge use cases.
“We are enabling our clients to build powerful, custom AI/ML pipelines without getting bogged down by the underlying technological complexities.” says Fausto Di Segni, Head of IoT and AI at SECO. “By integrating the Application Hub, the instant Clea trial, and the comprehensive documentation in our Developer Center, we are giving anyone building a smart product the power to not only test our AI, but to apply it to their own data streams”.
