Announcing Unreal Robotics Lab (URLab) -- MuJoCo Physics in Unreal Engine 5

Created on April 07, 2026

2026   ·   Unreal-Robotics-Lab   URLab   UnrealRoboticsLab   MuJoCo   Unreal-Engine   UE5   robotics   simulation   sim2real   synthetic-data   reinforcement-learning   ROS2   ·   robotics   simulation   open-source

I’m excited to share the alpha release of Unreal Robotics Lab (URLab) – an open-source Unreal Engine 5 plugin that embeds the MuJoCo physics engine directly into the editor and runtime.

TL;DR. URLab combines Unreal Engine’s photorealistic rendering with MuJoCo’s precision physics, targeting sim-to-real transfer, synthetic data generation, and robot learning workflows that need both visual fidelity and accurate contact dynamics.

Why URLab

Most robotics simulators force a trade-off: fast, accurate physics or photorealistic rendering. URLab removes that trade-off by running MuJoCo on an async physics thread inside Unreal Engine 5, with a component-based architecture that maps 1:1 to MuJoCo elements. You get the full MuJoCo C API from C++ and Blueprints, MJCF drag-and-drop import, and ZMQ networking to drive the sim from Python or ROS 2 – all inside Unreal’s editor.

What’s in the alpha

  • MJCF drag-and-drop import – drop a MuJoCo .xml into the Content Browser and URLab builds a full body / joint / actuator / sensor tree as an Articulation Blueprint.
  • Quick Convert – attach a component to any Static Mesh and it becomes a MuJoCo physics body. No XML required.
  • 40+ sensors, 8 actuator types, 4 joint types – covering position, velocity, motor, muscle, damper, adhesion, cylinder, general actuators; hinge, slide, ball, free joints; accelerometer, gyro, force/torque, touch, rangefinder, frame tracking, cameras.
  • ZMQ networking – stream sensor data out and receive actuator commands in via PUB/SUB sockets.
  • Python policy bridge (urlab_bridge) – remote control, RL policy deployment, sensor monitoring, ROS 2 bridging.
  • Record and replay, keyframe system, possess-and-walk control, cinematic tools, CoACD convex decomposition, Unreal Landscape -> MuJoCo heightfield conversion.
  • MjSimulate dashboard – in-editor widget with physics tuning, actuator sliders, live sensor readouts and camera feeds.

When to use it

  • You need photorealistic rendering with accurate contact physics – sim-to-real transfer, synthetic data generation for perception models.
  • You want to control MuJoCo robots from Python or ROS 2 while rendering in Unreal Engine.
  • You want Unreal’s ecosystem (Blueprints, Sequencer, Marketplace assets, Niagara) paired with physics-accurate robots.

Getting started

cd YourUnrealProject/Plugins
git clone https://github.com/URLab-Sim/UnrealRoboticsLab.git
cd UnrealRoboticsLab/third_party
./build_all.ps1

Regenerate project files, build, launch the editor, and drop an MJCF into the Content Browser. Full walkthrough in the Getting Started guide.

Requirements: Unreal Engine 5.7+, Windows (Linux experimental), MuJoCo 3.7+ (bundled), Visual Studio 2022, CMake 3.24+.

Citing URLab

If URLab supports your research, please cite the ICRA 2026 paper:

@inproceedings{embleyriches2026urlab,
  title     = {Unreal Robotics Lab: A High-Fidelity Robotics Simulator with Advanced Physics and Rendering},
  author    = {Embley-Riches, Jonathan and Liu, Jianwei and Julier, Simon and Kanoulas, Dimitrios},
  booktitle = {IEEE International Conference on Robotics and Automation (ICRA)},
  year      = {2026},
  url       = {https://arxiv.org/abs/2504.14135}
}

Getting involved

Issues, discussions, and contributions are welcome on GitHub. If you’re building on URLab or considering it for a project, I’d love to hear about it – reach out via the contact links on the homepage or drop a note in the repo’s Discussions.

Unreal Robotics Lab is an independent software plugin. It is NOT affiliated with, endorsed by, or sponsored by Epic Games, Inc. “Unreal” and “Unreal Engine” are trademarks of Epic Games, Inc.

References