Wendy LogoWendy
Back to Blog

Wendy Early Preview: Build, Deploy, and Tinker at the Edge

Wendy is now available as an early preview for developers, students, and tinkerers. Deploy containerized apps to Raspberry Pi, Jetson, or any Linux device with a single command.

Wendy LabsWendy Labs Team
April 3, 2026

We built Wendy because deploying to a robot or a Raspberry Pi shouldn't feel like a trip back to 2003. If you can ship a mobile app with one click, why should the physical world be any different?

Today, we're opening up Wendy as an early preview for developers, students, and tinkerers. This isn't a polished 1.0 — it's an invitation to build with us.

What You Can Do Today

Wendy gives you a complete workflow for deploying containerized apps to edge devices — Raspberry Pi, NVIDIA Jetson, or any Linux machine running the Wendy Agent.

wendy init
wendy run

wendy init walks you through an interactive wizard — pick your language, choose your entitlements (GPU, Bluetooth, persistent storage, etc.), and you're scaffolded. wendy run builds a container image, pushes it to your device, and streams logs back to your terminal. That's the whole workflow.

  • GPU-accelerated workloads on Jetson — run LLMs, YOLOv8, and vision models at the edge
  • Remote debugging with breakpoints, logging, and hot reloading from your IDE
  • Bluetooth, USB, WiFi, audio, and persistent storage through a simple entitlements model in wendy.json
  • Automatic device discovery via mDNS — no IP addresses to memorize
  • Your language, your choice — Python, Swift, Rust, TypeScript, C++

Whether you're wiring up a home automation system, running inference on a Jetson, building a robot, or just learning how containers work on real hardware — this is for you.

An Honest Note on Security

We believe in being upfront. Right now, Wendy has no authentication, encryption, or access control. Anyone on your network can SSH into a WendyOS device, and any Wendy CLI can deploy to any reachable agent. We built it this way deliberately to eliminate friction while we get the core developer experience right.

What this means for you:

  • Treat Wendy as a lab and learning tool — not a production system
  • Don't run it on untrusted or public networks
  • Don't use it to deploy anything sensitive
  • It's perfect for your desk, your home lab, your classroom, your hackathon, your garage

What's actively in development:

  • Project-scoped access control — lock down which users and CLI instances can deploy to which devices
  • End-to-end mTLS encryption — encrypted communication between CLI and agent
  • Production-grade device identity — secure onboarding and fleet management

We're building Wendy to take you from tinkering on your desk to running in the field. The security and access control to make that safe is coming shortly. We'll make the transition seamless.

For more details, see our Security documentation.

What We Need From You

This early preview exists because we want real feedback from real developers building real things — before we harden everything down.

  • Use it. Deploy something. Break something.
  • Tell us what's missing. File an issue on GitHub or jump into our Discord.
  • Share what you're building. We'd love to feature your projects.

The best tools are shaped by the people who use them. Help us shape Wendy.

Get Started

Ready to try it? Our Getting Started guide walks you through everything — installing the CLI, setting up your device, and deploying your first app. You'll be up and running in minutes.

Welcome to the early days. Let's build something.