Overview

PyMechanical is a Python interface for Ansys Mechanical (FEA software for structural engineering), enabling automation and integration of simulation workflows from 2024 R2 and later.

PyMechanical provides two distinct modes of interaction:

• Embedding mode — Run Mechanical directly in your Python process via the App class. Full object-model access, fast startup, ideal for Jupyter notebooks and interactive scripting.

• Remote session mode — Launch Mechanical as a separate server process and communicate via gRPC. Process isolation, optional GUI, ideal for CI/CD, Docker, and automation.

Installation

Install from PyPI:

pip install ansys-mechanical-core

For graphics support:

pip install ansys-mechanical-core[graphics]

For RPC functionality:

pip install ansys-mechanical-core[rpc]

For AI assistant integration via the Model Context Protocol (MCP):

pip install ansys-mechanical-core[mcp]

Requirements:

• Licensed copy of Ansys Mechanical (2024 R2 or later).

• For embedded mode: Local Mechanical installation required.

• For remote session mode: Network access to a running Mechanical instance.

Quick start

Embedding mode:

from ansys.mechanical.core import App

app = App(globals=globals())
print(app)

# Access Mechanical objects directly
Model.AddStaticStructuralAnalysis()

Remote session mode:

from ansys.mechanical.core import launch_mechanical

mechanical = launch_mechanical()
result = mechanical.run_python_script("2+3")
print(result)  # Output: 5
mechanical.exit()

Connect to existing instance:

from ansys.mechanical.core import connect_to_mechanical

mechanical = connect_to_mechanical(port=10000)
result = mechanical.run_python_script("Model.Name")
print(result)

Troubleshooting

• Connection refused: Ensure Mechanical is running and the port is accessible.

• License error: Verify your Ansys license is properly configured.

• Import error: Check that ansys-pythonnet is installed (not pythonnet).

• Linux embedding: Use mechanical-env to run Python scripts for embedded mode.

Documentation and support

• Documentation

• Cheat sheet

• PyMechanical-MCP — AI assistant integration via MCP

• Issues — report bugs or request features

• Discussions — ask questions

• Contributing guide