About FEA-MCP Server
FEA-MCP Server is an MCP server giving AI access to finite element analysis software, with a unified interface for ETABS and LUSAS modelling. The README states that the project is superseded by LUSAS-MCP-Server and is no longer actively developed.
Key features
- 01Unified interface for ETABS and LUSAS
- 02Batch creation of points, lines, surfaces and volumes
- 03Retrieval of modelled geometric objects
- 04Reading of model units
- 05Sweep and object selection in LUSAS
Use cases
- 01AI-assisted generation of structural geometry models
- 02Querying geometric objects in ETABS or LUSAS models
- 03Creating lines, surfaces and volumes by sweeping in LUSAS
MCP support
This project is an MCP server that exposes tools such as get_units, create_objects_by_coordinates and get_all_geometries to MCP clients, with available functions depending on the target software.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEWarning
This repository has been superseded
This project is no longer actively developed. Please use LUSAS-MCP-Server instead.
The newer project takes a different approach and provides full LUSAS API support.
![]()
FEA-MCP Server
An Finite Element Analysis Model Context Protocol Server for AI
🚀 Overview
The FEA-MCP Server provides a unified API interface for interacting with various Finite Element Analysis (FEA) software packages. It enables AI control of FEA modelling, analysis, and post-processing through a consistent interface, regardless of the underlying software implementation (currently ETABS and LUSAS are supported).
✨ Supported Features
- Multiple Software: Supports mainstream FEA software including:
- Geometric Modelling:
- Create Point/Joint
- Create Line/Frame/Beam/Column
- Create Volume/Solid
- Sweep Points/Lines/Surfaces (LUSAS only)
- Get modelled Points/Lines/Surfaces/Volumes
- Select objects (LUSAS only)
- Other:
- Read model units
🖥️ MCP Tools
The server provides the following main API functions:
get_units: Returns the model unitscreate_objects_by_coordinates: Batch-creates various geometric objects (points, lines/frames, surfaces/areas, volumes/solids)get_all_geometries: Returns all the modelled geometric objects (points, lines/frames, surfaces/areas, volumes/solids)get_points: Returns all the modelled points
(the following are only available for ETABS)
get_frames: Returns all the modelled framesget_areas: Returns all the modelled areas
(the following are only available for LUSAS)
get_lines: Returns all the modelled linesget_surfaces: Returns all the modelled surfacesget_volumes: Returns all the modelled volumessweep_points: Sweeps points to create linessweep_lines: Sweeps lines to create surfacessweep_surfaces: Sweeps surfaces to create volumesselect: Select modelled objects
🎯 Future Work
- Model Management: Define materials, sections, loads, and boundary conditions
- Analysis Control: Run simulations and retrieve results
- Coordinate System Support: Work with multiple coordinate systems
⚙️ Installation
Requirements
Required python libraries:
pywin32>=228 # Windows COM interface support
comtypes>=1.4.0 # Windows COM interface support
mcp>=0.1.0 # Model Control Protocol library
System Requirements:
- Windows operating system
- Installed FEA software (ETABS, LUSAS)
Guide
-
Install the required python libraries from command line:
pip install pywin32 comtypes mcp //or pip install -r requirements. txt -
Download this repository and save the extracted files locally (e.g. at
C:\your_path_to_the_extracted_server\FEA-MCP\). -
(Optional) Edit the MCP server configuration file, located at
src/config.json(see configuration section). By default the server is set to use LUSAS v21.1. -
Install Claude Desktop (or other AI client with MCP support).
-
Configure Claude Desktop to launch the MCP Server automatically (see Claude Desktop section).
-
You are good to go!
Configuration
The configuration file is located at src/config.json and contains the following main settings:
{
"server": {
"name": "FEA MCP",
"version": "1.0.0"
},
"fea": {
"software": "LUSAS",
"version": "21.1"
}
}
- server: Server name and version information
- fea:
software: FEA software (ETABS, LUSAS)version: software version (e.g. 21.1 for LUSAS)
🤖 AI Clients
5ire
Open 5ire > Tools > New, input the following info and then click Save:
| Input | Value |
|---|---|
| Tool Key | fea |
| Description | Finite Elements Analysis connection server (ETABS, LUSAS) |
| Command | python C:\your_path_to_the_extracted_server\FEA-MCP\src\server.py |
Caution: update the path! (single slashes)
Then turn on the server and you are good to go!
Claude Desktop
Open Claude Desktop and navigate to File > Settings > Developer > Edit Config, edit claude_desktop_config.json and add the following JSON.
{
"mcpServers": {
"fea": {
"command": "python",
"args": [
"C:\\your_path_to_the_extracted_server\\FEA-MCP\\src\\server.py"
]
}
}
}
Caution: update the path and use double backslash! Then restart Claude Desktop (from the tray icon, right click > Quit).
Visual Studio Code - GitHub Copilot
The MCP server is already setup in .vscode\mcp.json. Open the repository folder in visual studio, select .vscode\mcp.json and click "Start" above the server name (more info here).
🐞 Bugs
Google's Gemini seems to crash when the create_objects_by_coordinates tool is enabled.
🐛 Debug
You can test and debug the MCP server using the MCP Inspector by running the following command:
npx @modelcontextprotocol/inspector mcp run src\server.py
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.