Kit-less Installation

Kit-less Installation#

Isaac Lab can be installed and used without Isaac Sim using the kit-less mode. This is the fastest way to get started and is ideal for users who only need the Newton physics backend.

Preparing a Python Environment#

Creating a dedicated Python environment is strongly recommended. It helps:

  • Avoid conflicts with system Python or other projects installed on your machine.

  • Keep dependencies isolated, so that package upgrades or experiments in other projects do not break Isaac Sim.

  • Easily manage multiple environments for setups with different versions of dependencies.

  • Simplify reproducibility — the environment contains only the packages needed for the current project, making it easier to share setups with colleagues or run on different machines.

We recommend uv as the package manager — it is significantly faster than pip and conda for creating environments and resolving dependencies.

Caution

The Python version of the virtual environment must match the Python version of Isaac Sim.

  • For Isaac Sim 6.X, the required Python version is 3.12.

Using a different Python version will result in errors when running Isaac Lab.

The following instructions are for Isaac Sim 6.X, which requires Python 3.12.

  • Create a virtual environment using one of the package managers:

    uv is the fastest and recommended way to create a Python environment for Isaac Lab. To install uv, please follow the instructions here.

    # create a virtual environment named env_isaaclab with python3.12 and pip
uv venv --python 3.12 --seed env_isaaclab
# activate the virtual environment
source env_isaaclab/bin/activate

    :: create a virtual environment named env_isaaclab with python3.12 and pip
uv venv --python 3.12 --seed env_isaaclab
:: activate the virtual environment
env_isaaclab\Scripts\activate

    Note

    The --seed flag ensures pip is available inside the uv virtual environment, which is required by the Isaac Lab installer.

    To install conda, please follow the instructions here. You can create the Isaac Lab environment using the following commands.

    We recommend using Miniconda, since it is light-weight and resource-efficient environment management system.

    conda create -n env_isaaclab python=3.12
conda activate env_isaaclab

  • Ensure the latest pip version is installed. To update pip, run the following command from inside the virtual environment:

    uv pip install --upgrade pip

    uv pip install --upgrade pip

    Note

    If you are using pip directly instead of uv pip, replace uv pip with pip (or python -m pip on Windows) in the commands above and throughout this guide.

Cloning and installing Isaac Lab#

With the virtual environment activated, clone the repository:

git clone https://github.com/isaac-sim/IsaacLab.git
cd IsaacLab

Then install Isaac Lab (Newton backend, no Isaac Sim required) and kickoff training with MJWarp physics and the Newton visualizer:

# Install Isaac Lab (Newton backend, no Isaac Sim required)
./isaaclab.sh --install   # or ./isaaclab.sh -i

# Kickoff training with MJWarp physics and Newton visualizer
./isaaclab.sh -p scripts/reinforcement_learning/rsl_rl/train.py \
--task=Isaac-Cartpole-Direct-v0 \
--num_envs=16 --max_iterations=10 \
presets=newton_mjwarp --visualizer newton

:: Install Isaac Lab (Newton backend, no Isaac Sim required)
:: or: isaaclab.bat -i
isaaclab.bat --install

:: Kickoff training with MJWarp physics and Newton visualizer
isaaclab.bat -p scripts\reinforcement_learning\rsl_rl\train.py ^
--task=Isaac-Cartpole-Direct-v0 ^
--num_envs=16 --max_iterations=10 ^
presets=newton_mjwarp --visualizer newton

Features available in kit-less mode (Newton backend, no Isaac Sim):

  • Newton physics simulation (GPU-accelerated, including MuJoCo-Warp solver)

  • All manager-based and direct RL environments that support Newton

  • RL training with SKRL, RSL-RL, and other frameworks

  • Robot assets compatible with Newton

Features that require Isaac Sim:

  • PhysX physics backend

  • Isaac Sim RTX rendering (not ovrtx)

  • Kit visualizer

  • Photorealistic rendering workflows

  • ROS / ROS2 integration

  • URDF and MJCF importers (GUI-based)

  • Deformable objects and surface gripper (PhysX-only)

  • Teleoperation and imitation learning workflows

To install Isaac Sim, use the pip method described in Installation using Isaac Sim Pip Package.

Selective Install#

./isaaclab.sh -i always installs the full core package set (assets, tasks, physx, rl, visualizers, …). The argument controls which optional submodules and extra feature dependencies are added on top.

Optional submodules (heavy — must be explicitly requested):

Token

What it installs

mimic

isaaclab_mimic — imitation-learning tools (ipywidgets, h5py)

teleop

isaaclab_teleop — teleoperation SDK (Linux x86 only)

Optional extra feature sets (heavy deps on top of always-installed core):

Token

What it installs

newton

Newton physics library (newton[sim]) + newton extras across isaaclab_newton, isaaclab_physx, isaaclab_visualizers

rl[<framework>]

RL framework. Selectors: rsl-rl, skrl, sb3, rl-games. Omit selector for all.

visualizer[<backend>]

Visualizer backend. Selectors: rerun, viser, newton, kit. Omit selector for all.

contrib[rlinf]

rlinf extras (ray, diffusers, etc.)

ov

OVRTX + OVPhysX extras for Omniverse rendering

isaacsim

Isaac Sim pip package

all

Core + optional submodules (mimic, teleop) + auto extras (newton, rl, visualizer, ov). Default. Does not include contrib.

none

Core packages only — no optional submodules, no extra feature deps

Examples:

# Core only (physx, tasks, assets always included — no optional extras)
./isaaclab.sh -i none

# Newton physics + RSL-RL (most common kitless setup)
./isaaclab.sh -i newton,'rl[rsl-rl]'

# Newton + OVRTX renderer + RSL-RL + Newton visualizer
./isaaclab.sh -i newton,ov,'rl[rsl-rl]','visualizer[newton]'

:: Core only
isaaclab.bat -i none

:: Newton physics + RSL-RL
isaaclab.bat -i newton,rl[rsl-rl]

:: Newton + OVRTX + RSL-RL + Newton visualizer
isaaclab.bat -i newton,ov,rl[rsl-rl],visualizer[newton]

OVRTX Rendering#

OVRTX provides GPU-accelerated rendering for vision tasks without Kit.

./isaaclab.sh -i ov[ovrtx]

./isaaclab.sh -p scripts/benchmarks/benchmark_rsl_rl.py \
  --task Isaac-Repose-Cube-Shadow-Vision-Benchmark-Direct-v0 \
  --headless --enable_cameras --num_envs 16 --max_iterations 10 \
  presets=newton_mjwarp,ovrtx_renderer,simple_shading_diffuse_mdl

isaaclab.bat -i ov[ovrtx]

isaaclab.bat -p scripts\benchmarks\benchmark_rsl_rl.py ^
  --task Isaac-Repose-Cube-Shadow-Vision-Benchmark-Direct-v0 ^
  --headless --enable_cameras --num_envs 16 --max_iterations 10 ^
  presets=newton_mjwarp,ovrtx_renderer,simple_shading_diffuse_mdl

Running Installation Tests#

./isaaclab.sh -p -m pytest source/isaaclab/test/cli/test_cli_utils.py -v

isaaclab.bat -p -m pytest source\isaaclab\test\cli\test_cli_utils.py -v