Matterix

Matterix is a multi-scale, GPU-accelerated robotic simulation framework designed to create high-fidelity digital twins of chemistry labs, thus accelerating workflow development. This multi-scale digital twin simulates robotic physical manipulation, powder and liquid dynamics, device functionalities, heat transfer, and basic chemical reaction kinetics. This is enabled by integrating realistic physics simulation and photorealistic rendering with a modular GPU-accelerated semantics engine, which models logical states and continuous behaviors to simulate chemistry workflows across different levels of abstraction.

Key Features

The key features of Matterix are:

• Multi-scale simulation with integrated PhysX and semantics engines: supports robotics, soft bodies, liquids, powders, heat transfer, chemical kinetics, and device functions. Enables simulation of various vectorized logical events and continuous processes.
• Digital twin design: Easily create environments with multiple agents using an extensive asset library in wet lab domain.
• Workflow automation: Build and run experimental workflows with minimal effort.
• Real-world deployment: Deploy virtual workflows to physical robots and lab setups.

Installation

• Install Isaac Lab by following the installation guide. We recommend using the conda installation as it simplifies calling Python scripts from the terminal. To install IsaacSim and IsaacLab using conda:

# create and activate conda env
conda create -n <isaaclab-conda-env-name> python=3.11
conda activate <isaaclab-conda-env-name>
pip install --upgrade pip

# Install a CUDA-enabled PyTorch
pip install -U torch==2.7.0 torchvision==0.22.0 --index-url https://download.pytorch.org/whl/cu128
# Install the Isaac Lab packages along with Isaac Sim:
pip install isaaclab[isaacsim,all]==2.3.0 --extra-index-url https://pypi.nvidia.com

For advanced installation options, refer to installation guide.

• Install git lfs with git lfs install

• Clone this project/repository separately from the Isaac Lab installation (i.e. outside the IsaacLab directory):

  # Matterix has submodules
  git clone  --recurse-submodules https://github.com/ac-rad/Matterix.git
  cd Matterix
  git submodule foreach 'git lfs pull'

• If you are using the Isaac Sim python interpreter (i.e., not using Python venv or conda):

  # Ensure ${ISAACLAB_PATH} and ${ISAACSIM_PATH} are set in your terminal before running:
  ./matterix.sh -i

• If you are using conda that has Isaac Lab installed, install the library in editable mode using:

  # Activate your Isaac Lab Conda environment
  conda activate <isaaclab-conda-env-name>
  
  # Install Matterix packages in editable mode
  # This installs: matterix_sm, matterix_assets, matterix_tasks, and matterix
  # matterix_sm will auto-detect Isaac Lab and include full functionality
  python -m pip install -e source/*
  
  # ─── Fallback: If Isaac Lab is not installed in the active env ───
  # Ensure ${ISAACLAB_PATH} and ${ISAACSIM_PATH} are set in your terminal before running:
  ./matterix.sh -p -m pip install -e source/*

• If you are using Python venv that has Isaac Lab installed, install the library in editable mode using:

  # Activate your virtual environment
  source <path-to-your-venv>/bin/activate  # On Linux/macOS
  # .\<path-to-your-venv>\Scripts\activate  # On Windows
  
  # Install Matterix packages in editable mode
  python -m pip install -e source/*

Package Structure

MATteRIX consists of four installable packages:

1. matterix_sm - Standalone state machine for sequential action orchestration

   • Auto-detects Isaac Lab and installs full functionality when available
   • Falls back to minimal install (configs only) when Isaac Lab is not present
   • Can be used independently in other robotic projects

2. matterix_assets - Asset library with semantic metadata (robots, labware, infrastructure)

3. matterix_tasks - Task/environment definitions for RL training

4. matterix - Core simulation framework (depends on matterix_sm)

The command python -m pip install -e source/* installs all packages with optimal configuration.

Workflows

A workflow is a sequence of robotic actions orchestrated by the State Machine (SM) to accomplish tasks like picking, placing, or manipulating objects. The State Machine is hierarchical with arbitrary levels of abstraction: workflows contain compositional actions (like PickObject), which chain together primitive actions (Move, OpenGripper, CloseGripper). Primitive action policies are implemented in matterix_sm/primitive_actions/.

Workflows receive observations (including asset frame information streamed via the observation manager) as input and output action dictionaries that serve as input to the environment. Theoretically, a workflow is itself a compositional action and can be implemented using various planning methods for long-horizon tasks. The current state machine implementation is particularly useful for workflow testing and data collection.

Multi-Agent Support

Environments can contain multiple agents (robots, devices, etc.), each capable of executing actions. Every action must specify agent_assets - the agent name(s) responsible for executing that action:

• Single agent: agent_assets="robot"
• Multiple agents (joint action): agent_assets=["robot_1", "robot_2"]

Asset Frames

Asset manipulation frames (e.g., pre_grasp, grasp, post_grasp) are defined in body frame coordinates within asset configuration files in matterix_assets. These frames enable frame-based manipulation where actions can target specific object frames (e.g., MoveToFrameCfg(object="beaker", frame="grasp")).

Creating a Workflow

Define workflows in your environment config by creating action sequences:

from matterix_sm import PickObjectCfg, MoveToFrameCfg
from matterix_sm.robot_action_spaces import FRANKA_IK_ACTION_SPACE

# Define a workflow as a list of actions
workflows = {
    "pickup_beaker": [
        PickObjectCfg(
            object="beaker_500ml",
            agent_assets="robot",  # Which agent executes this action
            action_space_info=FRANKA_IK_ACTION_SPACE,
        ),
    ],
    "move_to_table": [
        MoveToFrameCfg(
            object="table",
            frame="center",  # Frame defined in asset config (body frame)
            agent_assets="robot",
            action_space_info=FRANKA_IK_ACTION_SPACE,
        ),
    ],
}

Add workflows to your environment config class as a workflows attribute. See source/matterix_sm/README.md for detailed action documentation and available primitive/compositional actions.

Running a Workflow

# List available workflows for a task
python scripts/list_workflows.py --task Matterix-Test-Beaker-Lift-Franka-v1

# Run a specific workflow
python scripts/run_workflow.py --task Matterix-Test-Beaker-Lift-Franka-v1 --workflow pickup_beaker --num_envs 4

# Or using matterix.sh wrapper
./matterix.sh -p scripts/run_workflow.py --task Matterix-Test-Beaker-Lift-Franka-v1 --workflow pickup_beaker

The State Machine automatically handles parallel execution across multiple environments, with each environment progressing through actions independently.

Verification

• Verify that the extension is correctly installed by:

  • Listing the available tasks:

    Note: It the task name changes, it may be necessary to update the search pattern "Template-" (in the scripts/list_envs.py file) so that it can be listed.

    # use 'FULL_PATH_TO_isaaclab.sh|bat -p' instead of 'python' if Isaac Lab is not installed in Python venv or conda
    python scripts/list_envs.py

  • Running a task:

    # use 'FULL_PATH_TO_isaaclab.sh|bat -p' instead of 'python' if Isaac Lab is not installed in Python venv or conda
    python scripts/<RL_LIBRARY>/train.py --task=<TASK_NAME>

  • Running a task with dummy agents:

    These include dummy agents that output zero or random agents. They are useful to ensure that the environments are configured correctly.

    • Zero-action agent

      # use 'FULL_PATH_TO_isaaclab.sh|bat -p' instead of 'python' if Isaac Lab is not installed in Python venv or conda
      python scripts/zero_agent.py --task=<TASK_NAME>
      
      # example of an environment with two robot arms and several beakers:
      python scripts/zero_agent.py --task Matterix-Test-Beakers-Franka-v1 --num_envs 1
      # or
      ./matterix.sh -p scripts/zero_agent.py --task Matterix-Test-Beakers-Franka-v1 --num_envs 1
      

    • Random-action agent

      # use 'FULL_PATH_TO_isaaclab.sh|bat -p' instead of 'python' if Isaac Lab is not installed in Python venv or conda
      python scripts/random_agent.py --task=<TASK_NAME>
      
      # example of an environment with two robot arms and several beakers:
      python scripts/random_agent.py --task Matterix-Test-Beakers-Franka-v1 --num_envs 1
      # or
      ./matterix.sh -p scripts/random_agent.py --task Matterix-Test-Beakers-Franka-v1 --num_envs 1
      

Set up IDE (Optional)

To setup the IDE, please follow these instructions:

• Run VSCode Tasks, by pressing Ctrl+Shift+P, selecting Tasks: Run Task and running the setup_python_env in the drop down menu. When running this task, you will be prompted to (1) add the absolute path to your Isaac Sim and (1) add the absolute path to your Isaac Lab installation.

If everything executes correctly, settings.json (and launch.json if missing) will be created in the .vscode directory. The file contains the python paths to all the extensions provided by Isaac Sim, Isaac Lab, and Omniverse. This helps in indexing all the python modules for intelligent suggestions while writing code.

Setup as Omniverse Extension (Optional)

We provide an example UI extension that will load upon enabling your extension defined in source/matterix/matterix/ui_extension_example.py.

To enable your extension, follow these steps:

1. Add the search path of this project/repository to the extension manager:

   • Navigate to the extension manager using Window -> Extensions.
   • Click on the Hamburger Icon, then go to Settings.
   • In the Extension Search Paths, enter the absolute path to the source directory of this project/repository.
   • If not already present, in the Extension Search Paths, enter the path that leads to Isaac Lab's extension directory directory (IsaacLab/source)
   • Click on the Hamburger Icon, then click Refresh.

2. Search and enable your extension:

   • Find your extension under the Third Party category.
   • Toggle it to enable your extension.

Code formatting

We have a pre-commit template to automatically format your code. To install pre-commit:

pip install pre-commit

Then you can run pre-commit with:

pre-commit run --all-files

Troubleshooting

Pylance Missing Indexing of Extensions

In some VsCode versions, the indexing of part of the extensions is missing. In this case, add the path to your extension in .vscode/settings.json under the key "python.analysis.extraPaths".

{
    "python.analysis.extraPaths": [
        "<path-to-ext-repo>/source/Matterix"
    ]
}

Pylance Crash

If you encounter a crash in pylance, it is probable that too many files are indexed and you run out of memory. A possible solution is to exclude some of omniverse packages that are not used in your project. To do so, modify .vscode/settings.json and comment out packages under the key "python.analysis.extraPaths" Some examples of packages that can likely be excluded are:

"<path-to-isaac-sim>/extscache/omni.anim.*"         // Animation packages
"<path-to-isaac-sim>/extscache/omni.kit.*"          // Kit UI tools
"<path-to-isaac-sim>/extscache/omni.graph.*"        // Graph UI tools
"<path-to-isaac-sim>/extscache/omni.services.*"     // Services tools
...