Edge OS for factories

Factory AI is moving onto the factory floor itself, and Wendy OS is one of the new software stacks being pitched as the layer to run it locally on edge devices. (iottechnews.com) Edge AI means running a model on or near the machine that produces the data, not in a distant cloud. IndustryWeek says that setup is used for jobs like defect detection, safety monitoring, and machine adjustments that need responses in milliseconds. (industryweek.com) That is the problem Wendy Labs says it wants to solve. Its GitHub page describes WendyOS as a custom Linux distribution based on Yocto and OpenEmbedded, built for devices such as NVIDIA Jetson and Raspberry Pi, with Docker, Secure Shell access, and over-the-air updates already wired in. (github.com) IoT Tech News reported last week that Wendy Labs introduced Wendy OS as an open-source platform for “physical AI” in manufacturing and said it aims to cut deployment timelines from months to minutes. The same report said factory teams often spend hundreds of engineering hours turning a proof-of-concept model into something that runs across fleets of robots, cameras, and sensors. (iottechnews.com) The pitch lands at a moment when larger vendors are making similar moves. Microsoft said on April 16 that its Foundry Local preview is designed for industrial settings where inference must run “at the machine level” on a server on the factory floor, inside an electrical cabinet, or at a remote plant without cloud connectivity. (microsoft.com) The common issue is reliability. Microsoft’s post says local deployments are aimed at disconnected and highly regulated sites, while Wendy’s materials emphasize remote device management and autonomous operation for hardware that cannot depend on a constant internet link. (microsoft.com) (github.com) (iottechnews.com) Wendy’s GitHub also shows what “OS-level stack” means in practice. The project bundles an app manager called `wendy-agent`, supports Swift, Python, Rust, and TypeScript, and includes tooling for remote debugging and container-based deployment to ARM hardware. (github.com) That makes the operating system less like a bare Linux image and more like a managed base layer for factory devices. For manufacturers testing computer vision, predictive maintenance, or robot control, the bet is that standardizing that layer will make local AI easier to ship and maintain across many sites. (github.com) (industryweek.com) The open question is whether smaller platforms like Wendy can win trust against bigger industrial and cloud vendors. But the direction is clear in the latest product launches: more factory AI is being packaged to run where the work happens, not somewhere else. (microsoft.com) (iottechnews.com)