Warning
Pre-Alpha Software — Not an Early Access or General Availability Release.
Isaac Lab Arena v0.1.x is an early code release intended to give the community a practical starting point to experiment, provide feedback, and influence future design direction. APIs are unstable and will change. Features are incomplete. Documentation is evolving. Do not use this in production. See Project Status for details.
Isaac Lab Arena is an open-source extension to NVIDIA Isaac Lab that simplifies the creation of large-scale task and environment libraries for robotic policy evaluation. Co-developed with Lightwheel, it provides a composable architecture where environments are assembled on-the-fly from independent, reusable building blocks — eliminating the redundant boilerplate that plagues traditional task library development.
Instead of hand-writing and maintaining a separate configuration for every combination of robot, object, and scenario, Arena lets you compose environments from three independent primitives:
| Primitive | Description |
|---|---|
| Scene | The physical environment layout — a collection of objects, furniture, fixtures |
| Embodiment | The robot and its observations, actions, sensors, and controllers |
| Task | The objective — what the robot should accomplish (pick-and-place, open door, etc.) |
The ArenaEnvBuilder composes these primitives into a standard ManagerBasedRLEnvCfg that runs natively in Isaac Lab.
With the rise of generalist robot policies (e.g., GR00T N, pi0, SmolVLA), there is an urgent need to evaluate these policies across many diverse tasks and environments. Traditional approaches suffer from:
- Code duplication — each task variation (different object, different robot) requires a near-copy of the same configuration
- Maintenance burden — N robots × M objects × K scenes = an explosion of configs to keep in sync
- Slow iteration — researchers spend more time wrangling configs than running experiments
Arena solves this by making environment variation a first-class concept. Swap an object, change a robot, or modify a scene — all without duplicating a single line of task logic.
- Composable Environments — Mix and match scenes, embodiments, and tasks independently
- On-the-fly Assembly — Environments are built at runtime; no duplicate config files to maintain
- Asset Registry — Centralized management of robots, objects, and scenes with affordance annotations
- Integrated Evaluation — Built-in metrics and evaluation pipelines for policy benchmarking
- Teleoperation Support — Data collection via keyboard, VR, or other input devices
- GR00T Integration — First-class support for NVIDIA GR00T N policy training and evaluation
- LeRobot Hub — Publish and share environments on the Hugging Face LeRobot Environment Hub
- Linux (Ubuntu 22.04+)
- NVIDIA GPU (see Isaac Sim hardware requirements)
- Docker and NVIDIA Container Toolkit
- Git
Isaac Lab Arena currently supports installation from source inside a Docker container.
# 1. Clone the repository
git clone git@github.com:isaac-sim/IsaacLab-Arena.git
cd IsaacLab-Arena
git submodule update --init --recursive
# 2. Launch the Docker container
# Base container (recommended for development):
./docker/run_docker.sh
# Or with GR00T dependencies (for policy training/evaluation):
./docker/run_docker.sh -g
# 3. Verify the installation
/isaac-sim/python.sh -m pytest -sv -m "not with_cameras" isaaclab_arena/tests/Note: The Docker script automatically mounts
$HOME/datasets,$HOME/models, and$HOME/evalfrom your host into the container.
For detailed setup instructions (including server-client mode for GR00T), see the Installation Guide.
Compose a Franka arm in a kitchen scene with a couple of objects:
from isaaclab_arena.assets.asset_registry import AssetRegistry
from isaaclab_arena.cli.isaaclab_arena_cli import get_isaaclab_arena_cli_parser
from isaaclab_arena.environments.arena_env_builder import ArenaEnvBuilder
from isaaclab_arena.environments.isaaclab_arena_environment import IsaacLabArenaEnvironment
from isaaclab_arena.scene.scene import Scene
asset_registry = AssetRegistry()
# Select building blocks
background = asset_registry.get_asset_by_name("kitchen")()
embodiment = asset_registry.get_asset_by_name("franka")()
cracker_box = asset_registry.get_asset_by_name("cracker_box")()
tomato_soup_can = asset_registry.get_asset_by_name("tomato_soup_can")()
# Compose the environment
scene = Scene(assets=[background, cracker_box, tomato_soup_can])
env_cfg = IsaacLabArenaEnvironment(
name="franka_kitchen_example",
embodiment=embodiment,
scene=scene,
)
args_cli = get_isaaclab_arena_cli_parser().parse_args([])
env_builder = ArenaEnvBuilder(env_cfg, args_cli)
env = env_builder.make_registered()
env.reset()Explore more examples in the documentation, including:
| Example | Description |
|---|---|
| G1 Loco-Manipulation | Unitree G1 humanoid locomotion + manipulation |
| GR1 Open Microwave Door | GR1 humanoid interacting with articulated objects |
| GR1 Sequential Pick & Place | Multi-step manipulation with GR1 |
| Franka Lift Object | Classic Franka pick-up task |
IsaacLab-Arena/
├── isaaclab_arena/ # Core framework (environments, tasks, scenes, embodiments)
├── isaaclab_arena_environments/ # Concrete environment definitions
├── isaaclab_arena_examples/ # Policy and relation examples
├── isaaclab_arena_g1/ # Unitree G1 humanoid embodiment + examples
├── isaaclab_arena_gr00t/ # GR00T policy integration
├── docker/ # Docker configurations and launch scripts
├── docs/ # Sphinx documentation source
├── osmo/ # Cloud deployment configs (OSMO)
├── submodules/ # Git submodules (Isaac Lab, etc.)
├── setup.py # Package installation
├── CONTRIBUTING.md # Contribution guidelines
└── LICENSE.md # Apache 2.0 license
| Isaac Lab Arena | Isaac Lab | Isaac Sim | Python |
|---|---|---|---|
main branch |
2.3.2 | 5.1.0 | ≥ 3.10 |
release/0.1.1 |
2.3.0 | 5.0.0 | ≥ 3.10 |
Isaac Lab Arena is in pre-alpha (v0.1.x). This is important to understand:
| What This Means | Details |
|---|---|
| Not EA / GA | This is not an Early Access or General Availability release. It is a very early community code drop. |
| APIs will break | Public interfaces are under active development and will change without deprecation warnings. |
| Features are incomplete | Core capabilities like natural-language object placement, composite task chaining, RL task setup, and heterogeneous parallel evaluation are planned but not yet implemented. |
| Docker-only install | Source installation in a Docker container is the only supported method in v0.1.x. |
| Limited testing | The main branch contains the latest code but may not be fully tested. Use release/0.1.1 for the most stable experience. |
Isaac Lab Arena is part of a growing ecosystem of tools and benchmarks:
- Lightwheel RoboCasa Tasks — 250+ open-source tasks built on Arena
- Lightwheel LIBERO Tasks — Adapted LIBERO benchmarks
- LeRobot Environment Hub — Share and discover Arena environments on Hugging Face
- RoboTwin 2.0 — Extended simulation benchmarks using Arena
- Isaac Lab Teleop — Demonstration collection
- Isaac Lab Mimic — Synthetic data generation
We welcome contributions — bug reports, feature suggestions, and code. This is a pre-alpha project, so community input directly shapes the framework's direction.
- Read the Contribution Guidelines
- Sign off your commits (DCO required — see
CONTRIBUTING.md) - Open a Pull Request
Areas where contributions are especially valuable:
- New task definitions and benchmark suites
- Additional robot embodiments and scene assets
- Sim-to-real validated evaluation methods
- Documentation improvements and tutorials
- Questions & Ideas — GitHub Discussions
- Bug Reports — GitHub Issues
- Isaac Sim Questions — NVIDIA Forums
- Community Chat — Omniverse Discord
Isaac Lab Arena is released under the Apache 2.0 License.
Note that Isaac Lab Arena requires Isaac Sim, which includes components under proprietary licensing terms. See the Isaac Sim license for details.
If you use Isaac Lab Arena in your research, please cite:
@misc{isaaclab-arena2025,
title = {Isaac Lab Arena: Composable Environment Creation and Policy Evaluation for Robotics},
author = {{NVIDIA Isaac Lab Arena Contributors}},
year = {2025},
url = {https://github.com/isaac-sim/IsaacLab-Arena}
}If you use Isaac Lab (the underlying framework), please also cite the Isaac Lab paper.
Isaac Lab Arena is co-developed with Lightwheel and builds on NVIDIA Isaac Lab. We thank the Isaac Lab team and the broader robotics community for their foundational work.
Isaac Lab Arena · Pre-Alpha · Documentation · GitHub
Made with ❤️ by the NVIDIA Robotics Team