An open-source, engine-independent simulator for drones, fixed-wing aircraft and ground vehicles, with optional Unreal Engine 5 rendering.
microsoft/AirSim
An open-source simulator built on Unreal Engine for drones, cars and other autonomous vehicles, with APIs for AI and computer vision research.
- Stars
- 18,602
- Forks
- 4,928
- Open issues
- 781
- Last commit
- 2 months ago
What is it?
AirSim is a simulator for drones, cars and other autonomous vehicles, built as a plugin for Unreal Engine that can be added to any Unreal environment. An experimental Unity version is also available. It supports software-in-the-loop simulation with flight controllers such as PX4 and ArduPilot, and hardware-in-the-loop simulation with PX4.
The project is aimed at AI research, letting developers experiment with deep learning, computer vision and reinforcement learning for autonomous vehicles. It exposes platform-independent APIs for retrieving sensor data and controlling vehicles, with clients available in C++, Python, C# and Java.
Besides programmatic control, AirSim includes manual driving with a remote control or arrow keys, a data recording mode for capturing pose and images, and a Computer Vision mode that collects depth, segmentation and surface normal images from free-moving cameras without physics.
Key features
- 01Unreal Engine plugin that can be dropped into existing environments
- 02Software-in-the-loop simulation with PX4 and ArduPilot
- 03Hardware-in-the-loop simulation with PX4
- 04Client APIs in C++, Python, C# and Java for retrieving data and controlling vehicles
- 05Computer Vision mode for capturing depth, segmentation and surface normal images
Use cases
- Training deep learning or reinforcement learning policies for drones and self-driving cars before real-world testing
- Generating labeled camera datasets with depth and segmentation ground truth
- Testing flight controller code in simulation before deploying to a physical vehicle
Quick start
- 1
Download precompiled binaries from the releases page
- 2
Or build it from source following the platform build guide
Things to know
- Runs as a plugin for Unreal Engine, so you need Unreal Engine set up to use it.
- macOS has build-from-source instructions only, with no precompiled binaries listed.
- The Unity plugin is experimental.
In collections
Topics
FAQ
- What is AirSim?
- An open-source simulator built on Unreal Engine for drones, cars and other autonomous vehicles, with APIs for AI and computer vision research. AirSim is a simulator for drones, cars and other autonomous vehicles, built as a plugin for Unreal Engine that can be added to any Unreal environment. An experimental Unity version is also available. It supports software-in-the-loop simulation with flight controllers such as PX4 and ArduPilot, and hardware-in-the-loop simulation with PX4.
- Is AirSim free to use? What license does it have?
- AirSim is open source under the Other license. Check the license terms before using it in commercial products.
- Is AirSim still maintained?
- Updated within 6 months. The last commit to AirSim was on Sep 15, 2026.
- What can I use AirSim for?
- Training deep learning or reinforcement learning policies for drones and self-driving cars before real-world testing Generating labeled camera datasets with depth and segmentation ground truth Testing flight controller code in simulation before deploying to a physical vehicle
- How do I install AirSim?
- Download precompiled binaries from the releases page Or build it from source following the platform build guide
- Where can I get AirSim?
- The source code and installation instructions are on GitHub: https://github.com/microsoft/AirSim
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Open-source simulator for autonomous driving research, providing urban environments, digital assets, and sensor simulation built on Unreal Engine.
A Unreal Engine-based simulator for drones, cars and other vehicles, extended by Cosys-Lab with new sensor models and APIs.
Open-source robot simulator for modeling, programming and simulating robots, vehicles and mechanical systems.
An Unreal Engine 5 plugin that runs MuJoCo physics inside the editor and runtime, pairing accurate contact dynamics with Unreal's rendering.
An open ROS2 framework to simulate and deploy perception-based drone swarms across PX4, ArduPilot and Betaflight autopilots.