Installation

Prerequisites

  • Python 3.11

  • A compiled MOOSE framework installation (phase_field-opt binary)

  • Node.js 18+ (for the frontend)

  • Git

  • (Optional, for HPC execution) a NERSC account with sshproxy configured — see Execution Backends

Clone the Repository

git clone https://github.com/sukritimanna/AutoMOOSE.git
cd AutoMOOSE

Python Dependencies

conda create -n automoose python=3.11
conda activate automoose
pip install -r requirements.txt

Environment Configuration

Copy the example config and edit it:

cp config.env.example config.env

Set the core values in config.env:

# execution backend
EXECUTION_BACKEND=local
MOOSE_EXEC=/path/to/your/moose/phase_field-opt
RUNS_DIR=./runs

# model backend (model-agnostic)
LLM_PROVIDER=anthropic
LLM_MODEL=claude-sonnet-4-6
# provider credentials as required, e.g. ANTHROPIC_API_KEY=...

For HPC execution, additionally set the HPC_* values described in Execution Backends.

Frontend Dependencies

cd frontend
npm install
cd ..

Verify Installation

export $(grep -v '^#' config.env | xargs)
python -c "from automoose.server import app; print('Backend OK')"