Why Copper
Already showing up in: ✈️ Flying | 🚗 Driving | 🌊 Swimming | 🚀 Spacefaring | 🤖 Humanoids
Try Copper In 30 Seconds
No setup required. Open one of the live demos in your browser: the simulator runs on the left and the live Copper monitor runs on the right.
These are not mockups: BalanceBot is the exact same application that runs on a Raspberry Pi physical robot, and Flight Controller is the same control stack we run on a microcontroller (STM32H7) on real drones. Copper lets that same graph be recompiled for embedded hardware, a local workstation, or the browser.
Prefer a native app instead of the browser? Install the published demos directly:
The source code for these demos lives in examples/cu_rp_balancebot and examples/cu_flight_controller.
Want to see more Copper in action? Watch the community showcase video.
Get Started
- Start a new project from templates: Project Templates
- See a full task graph + runtime walkthrough: Copper Application Overview
- Build and deploy an application: Build and Deploy a Copper Application
- RON configuration reference: Copper RON Configuration Reference
Documentation
Link to the full documentation
- Runtime concepts and SDK features: Copper Runtime Overview
- Task lifecycle: Task Lifecycle
- Modular configuration: Modular Configuration
- Task automation: Task Automation with just
- Supported platforms: Supported Platforms
- Bare-metal development: Baremetal Development
- Available components: Available Components
- FAQ: FAQ
- Release notes: Copper Release Notes
- Roadmap: Roadmap
Python Support
Copper has two very different Python stories:
- Offline Python log analysis: use
cu29-exportand app-specific PyO3 modules such as examples/cu_flight_controller. This is a reasonable workflow because Python stays off the runtime hot path. - Runtime Python task prototyping: use components/tasks/cu_python_task and examples/cu_python_task_demo. This is for experimentation only and is strongly not recommended for production or realtime robots.
Putting Python inside a Copper task defeats the performance model Copper is built for: it adds allocations, latency, jitter, and middleware overhead, and it ruins the realtime characteristics of the stack. The intended use is to sketch one task in Python, get the behavior right, then rewrite it in Rust.
Citation
If you use Copper-rs in your research, please cite it as:
Project
[!NOTE] We are looking for contributors to help us build the best robotics framework possible. If you are interested, please join us on Discord or open an issue.