Repolode
Submit
autorope

autorope/donkeycar

Open-source Python library and hardware platform for building small-scale self-driving RC cars, with a simulator for experimenting before you build one.

View on GitHubWebsite
Stars
3,537
Forks
1,363
Open issues
140
Last commit
3 days ago

What is it?

Donkeycar is a minimalist, modular Python library for building and training small self-driving robot cars. It targets hobbyists, students and researchers, and is built around fast experimentation and easy community contributions. A car is defined as a pipeline of parts that run in order on each pass through the vehicle loop, so you can swap in cameras, controllers, actuators or your own code without rewriting the whole system.

The platform supports three kinds of autopilot: deep learning (with TensorFlow, TensorFlow Lite and PyTorch), GPS path following, and computer vision. It also provides a web UI, a built-in simulator for trying out driving models before any hardware exists, and tools for recording camera images and driving telemetry.

Installation is set up for Raspberry Pi, macOS and Linux/Windows PCs using uv for environment management, with separate install extras for each platform.

Key features

  • 01Modular vehicle pipeline of parts that can be swapped or extended.
  • 02Three autopilot types: deep learning, GPS path following and computer vision.
  • 03Support for TensorFlow, TensorFlow Lite and PyTorch models.
  • 04Built-in simulator for testing driving before building hardware.
  • 05Web UI with browser-based game controller and phone touch controls.
  • 06Data recording of camera images, steering and throttle inputs.

Use cases

  • Building a small robot car on a Raspberry Pi that learns to drive itself.
  • Teaching self-driving and neural networks in high school or university courses.
  • Experimenting with autopilot models in the simulator before hardware is ready.
  • Competing in DIY robocar races, including online simulator races.

Quick start

  1. 1

    Install uv, the environment manager (run the official install script from the README)

  2. 2

    Create a virtual environment for a PC (Linux/Windows)

    uv venv ~/.venvs/donkeycar --python 3.12
  3. 3

    Activate the virtual environment

    source ~/.venvs/donkeycar/bin/activate
  4. 4

    Install Donkeycar from PyPI with the PC extras

    uv pip install donkeycar[pc]

Things to know

  • Raspberry Pi setup needs system packages (python3-libcamera, python3-picamera2) that are not on PyPI.
  • Driving a real car requires building hardware; the bundled simulator is the option without a robot.

In collections

Computer vision libraries & toolsRobotics & embodied AI projects

Topics

#self-driving#robotics#raspberry-pi#deep-learning#computer-vision#simulator#python#cv2#donkeycar#jetson-nano

FAQ

What is donkeycar?
Open-source Python library and hardware platform for building small-scale self-driving RC cars, with a simulator for experimenting before you build one. Donkeycar is a minimalist, modular Python library for building and training small self-driving robot cars. It targets hobbyists, students and researchers, and is built around fast experimentation and easy community contributions. A car is defined as a pipeline of parts that run in order on each pass through the vehicle loop, so you can swap in cameras, controllers, actuators or your own code without rewriting the whole system.
Is donkeycar free to use? What license does it have?
donkeycar is open source under the MIT license. Check the license terms before using it in commercial products.
Is donkeycar still maintained?
Actively maintained. The last commit to donkeycar was on Nov 4, 2026.
What can I use donkeycar for?
Building a small robot car on a Raspberry Pi that learns to drive itself. Teaching self-driving and neural networks in high school or university courses. Experimenting with autopilot models in the simulator before hardware is ready. Competing in DIY robocar races, including online simulator races.
How do I install donkeycar?
Install uv, the environment manager (run the official install script from the README) Create a virtual environment for a PC (Linux/Windows): uv venv ~/.venvs/donkeycar --python 3.12 Activate the virtual environment: source ~/.venvs/donkeycar/bin/activate Install Donkeycar from PyPI with the PC extras: uv pip install donkeycar[pc]
Where can I get donkeycar?
The source code and installation instructions are on GitHub: https://github.com/autorope/donkeycar

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects