Skip to content

Installation guide

For users wanting to install a stable release of amorphouspy, we recommend using conda-forge. The package is available on conda-forge via a multi-output recipe, allowing you to install either the core simulation workflow package, or the web API server with its advanced process orchestrators. Installing via conda automatically handles all complex dependencies including LAMMPS.

(Note: Users are free to use whatever tool they prefer to manage their conda/mamba/pixi environments here).

Core Package

To install the core amorphouspy library with LAMMPS and all numerical/scientific dependencies: 

conda install -c conda-forge amorphouspy

Web API Server

To install the web API & Model Context Protocol (MCP) server integration, which automatically pulls in the core amorphouspy library, executorlib, fastapi, and associated packages: 

conda install -c conda-forge amorphouspy-api

From git / source (for development)

If you want to contribute to development or run the latest source code of this project from git, we recommend using pixi for environment management as it handles all dependencies — including LAMMPS with OpenMPI and API dependencies — in a single step using the local clone of the repository.

# Clone the repository
git clone https://github.com/glasagent/amorphouspy.git
cd amorphouspy

# Install the environment (core package, LAMMPS, API dependencies, etc.)
pixi install

Note: The pixi environment installs LAMMPS with OpenMPI support (lammps =2024.08.29=*_openmpi_*), which provides the lmp_mpi executable required by all simulation workflows.

Using pip only (analysis-only)

If you only need the analysis and structure generation tools (no LAMMPS simulations):

pip install amorphouspy

Warning: Without a LAMMPS installation, the simulation workflows (melt_quench_simulation, md_simulation, elastic_simulation, etc.) will not work. Only structure generation and analysis functions will be available.