Overview
The open source operating system and toolchain for Physical AI — robots, drones, and edge devices.
Physical AI Platform
The Full Stack Physical AI Platform
Develop, deploy, debug, and observe physical AI apps from a local CLI and editor workflow.
Get Started
Install Wendy
Plug in an NVIDIA Jetson or Raspberry Pi over USB and start deploying in seconds. Wendy for macOS is available in beta for Apple Silicon targets.
Wendy CLI
Install this on your developer machine.
curl -fsSL https://install.wendy.dev/cli.sh | bashwendy-agent for Linux
OptionalInstall this on an existing Linux target. WendyOS images already include it.
curl -fsSL https://install.wendy.dev/agent.sh | bashDeveloper machine setup
Configure your local tools, editor, and first device connection.
Open setup guideSupported Hardware
Wendy runs on edge devices with strong local compute. Plug one in over USB-C and start deploying.
NVIDIA Jetson
Orin Nano, AGX Orin, AGX Thor
- Up to 2000 TOPS AI performance
- CUDA, PyTorch & MLX support
- AGX Thor USB recovery flashing
- Hardware video encode & decode
Raspberry Pi
Pi 3, 4 & 5 (8GB Pi 5 recommended)
- Low power consumption
- Broad GPIO ecosystem
- Hardware PWM, SPI & I2C
- Affordable entry point
NVIDIA Jetson AGX Thor
High-performance physical AI at the edge
- Up to 2000 TOPS AI performance
- CUDA, PyTorch & MLX support
- USB recovery flashing
- Hardware video encode & decode
Linux
Run Wendy on your Linux development machine
- Develop and deploy from your Linux machine
- Works with USB-C and LAN devices
- Full Wendy CLI support
- Ideal for robotics and edge AI workflows
Core Components
Wendy consists of several core components:
WendyOS
The Apache 2.0 open source Linux distribution optimized for Physical AI. Includes a mobile-like USB-C development experience, remote debugging, and atomic updates.
Wendy for macOS
Deploy native macOS apps to Apple Silicon Mac targets with Wendy for macOS beta. Supports SwiftPM and Xcode projects with the same CLI workflow.
Wendy Cloud
A fleet management solution for OTA updates, crash reporting, and remote telemetry. Built for teams scaling from prototypes to thousands of devices.
Developer Tools
CLI and VSCode extension for interactive breakpoints, hot reloading, and one-click deployment. Works with VSCode, Cursor, and Windsurf.
Wendy Frameworks
Performance-first Swift, Python, and Rust frameworks for running AI models, controlling I/O, and handling high-speed telemetry.
Wendy Lite
WASM-based runtime for microcontrollers. Deploy Swift, Rust, C, or Zig apps to ESP32-C6 with the same CLI workflow.