The open-source Wendy OS aims to scale physical AI across manufacturing edge networks in minutes, not months.
Behind the modern factory’s automated guided vehicles and vision sensors, plant managers routinely face configuration delays when connecting local machine intelligence to broader supply chain infrastructure. The process of taking a proof-of-concept AI model and deploying it across hundreds of robotic arms or logistics sensors requires hundreds of engineering hours.
This week, a new company named Wendy announced an OS designed to make creating physical AI systems incredibly easy, stating it can reduce deployment timelines from months to minutes.
Maximilian Alexander, Co-Founder of Wendy, said: “Wendy is an operating system and developer platform for physical AI—built to make it dramatically easier to build and deploy on NVIDIA Jetson, Raspberry Pi, and other edge devices.
“We think robotics, edge AI, industrial systems, autonomous machines, and smart cameras should be far simpler to create. Less setup. Less infrastructure pain. Faster time to first demo. This is the start of something big.”
The industrial sector frequently treats edge hardware, like Raspberry Pis and NVIDIA Jetsons, as standard headless servers. In practice, facility operators view these units as basic SSH targets with no independent identity or agency within the network. However, these devices function as active participants in the physical world, operating as robots, drones, cameras, and sensors.
Plant directors require these endpoints to act autonomously to maintain high-volume production lines. The software must accommodate a range of form factors, including humanoids, factory robots, autonomous vehicles, and satellite constellations. The creators of Wendy draw inspiration from the ‘Ghost in the Shell’ franchise, viewing the boundary between digital and physical forms as the space where consciousness emerges.
Engineering constraints vs plant floor requirements
For operations directors, maintaining tight alignment between hardware and software dictates facility yield and equipment uptime. Building reliable systems for physical AI usually requires a PhD in systems engineering.
By providing the Apache 2.0 open source WendyOS and toolchain, the platform allows small teams to ship projects without relying on bloated enterprise software stacks. The goal is to bring the development simplicity that iOS and Android brought to mobile applications straight into the domains of robotics and drones. This approach removes the configuration nightmares that delay pilot projects from reaching full production scale.
Maintaining remote edge hardware creates friction for maintenance teams managing distributed assets. Sending technicians to physically connect to sensors located on offshore oil rigs or regional logistics hubs consumes heavy portions of the operating budget. WendyOS addresses this by incorporating built-in Mender integration to facilitate reliable over-the-air updates backed by A/B partition redundancy. If an update fails halfway through deployment, the device reverts to the previous working state, preventing costly downtime.
The OS itself operates as a custom Linux distribution based on Yocto/OpenEmbedded, specifically optimised for edge computing devices. This technical foundation delivers a minimal, secure, and production-ready environment suitable for long-term IoT deployments. Under the hood, the system maintains a low overhead footprint using a headless, systemd-based init setup optimised for embedded hardware.
How Wendy OS is bridging IT workflows with edge infrastructure
Integrating cloud-native workflows into the factory floor presents ongoing compatibility issues for industrial IT departments attempting to scale their networks. Wendy bypasses legacy deployment methods by offering automatic Docker containerisation and multi-architecture builds.
Out of the box, the developer-friendly OS comes pre-configured with Docker, SSH, and essential development tools. The platform actively brings the ease of modern cloud development directly to the edge computing world.
The main development environment acts as a comprehensive CLI, enabling engineers to build, deploy, and debug applications directly on ARM-based hardware like the NVIDIA Jetson and Raspberry Pi.
The core agent, wendy-agent, operates as an app manager written in Swift. This tool handles application deployment and network configuration, maintaining support for both NetworkManager and ConnMan protocols.
Using simple commands like wendy run handles the underlying complexity of edge deployment. For engineers tasked with troubleshooting code on remote factory equipment, the toolkit provides full LLDB debugging support for edge applications. A dedicated Visual Studio Code extension, wendy-vscode, links the CLI directly to the device management workflow.
Powering physical AI: From GPUs to local sensors
Wendy is heavily hardware-optimised for edge devices, focusing on purpose-built support for the NVIDIA Jetson Orin Nano. The platform includes a Yocto meta-layer, named meta-wendyos-jetson, which establishes a Docker-based build environment and OTA update support specific to the developer kit.
The system accommodates varied engineering backgrounds by providing multi-language support encompassing Swift, Python, Rust, and TypeScript/Node.js. Developers can write code on macOS or Linux machines and deploy directly to ARM devices. The platform includes sample applications to demonstrate how to build projects across these different languages.
The platform supplies purpose-built system libraries to manage industrial communication and machine learning inference. Relying on Swift 6.2+ with modern async/await patterns, the toolkit processes complex data streams. It provides TensorRT Swift 6.2 bindings for Linux, enabling high-performance deep learning inference directly on NVIDIA hardware. DeepStream Swift bindings allow plant teams to build intelligent video analytics applications for defect detection and facility monitoring.
For interprocess communication across the Linux operating system, Wendy features a pure Swift 6 D-Bus protocol implementation, built with SwiftNIO and modern concurrency support. Hardware peripheral control relies on a Swift Bluetooth (BlueZ) library for Linux, allowing BLE communication from the edge devices to local sensors. Processing visual data feeds relies on a Swift 6.2 GStreamer wrapper aimed directly at robotics and computer vision use cases.
The creators note the Wendy logo visualises overlapping forms that symbolise the marriage of software and hardware, mind and body, ghost and shell. The stated intent is to move beyond deploying code to metal boxes and give these devices a soul—an operating system that lets them think, act, and evolve.
By removing the friction of manual configuration, this toolchain provides the industrial sector with a direct method for deploying physical AI at scale.
See also: Visual-Language-Action mechanisms in next-gen AI for IIoT
Want to learn more about the IoT from industry leaders? Check out IoT Tech Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including AI & Big Data Expo and the Cyber Security Expo. Click here for more information.
IoT News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.
