Back to Knowledge Base
454x
002032
Dipl.-Ing. Frank Faulstich
2026-01-22

Integration of RFEM 6 into AI Applications via MCP Server

As of the version 6.13.0002 (released on January 19, 2026), Dlubal Software provides a server based on the Model Context Protocol (MCP) for the first time. This interface allows for the integration of RFEM 6 as a tool in agentic AI systems and chatbots.

What is the Model Context Protocol (MCP)?

The Model Context Protocol is an open standard initiated by Anthropic that standardizes the connection between large language models (LLMs) and external data sources or software tools.

Functionality and Architecture

Until now, integrations between AI models and specialized applications often required individual, proprietary implementations for each combination of model and tool. MCP solves this problem with a standardized architecture based on JSON-RPC 2.0 messages.

The architecture is divided into three roles:

  1. MCP Host: The parent application (for example, Claude Desktop, IDEs such as Cursor, or your own chatbot) that controls the interaction.
  2. MCP Client: The component within the host that maintains the connection to the server.
  3. MCP Server: The program that provides specific features or data. In the context of Dlubal, the new RFEM 6 MCP server takes on this role.

Official information and the complete specification of the protocol can be found at:

Significance for Development of AI Agents

The main difference to conventional chatbot integrations lies in the system's ability to act. While classic RAG (Retrieval-Augmented Generation) systems primarily retrieve information from documents, MCP allows the use of “tools.”

For structural design, this means that an LLM no longer just talks about static problems, but can actively interact using RFEM. The AI can now do everything that is possible via the API:

  • Create or modify model geometries.
  • Define load cases and start calculations.
  • Read results.

This standardization eliminates the need for developers to redefine complex API logic for each AI model, as the LLM can dynamically discover and use the available tools on the server.

Running MCP Server in RFEM 6

The MCP server is part of the standard installation of RFEM 6.13.0002 and does not require a separate download.

Start via Command Line

To do this, open a PowerShell and navigate to the RFEM 6 installation directory (by default C:\Program Files\Dlubal\RFEM 6.13\bin\).

Start the server using this command: 

.\python.exe -m rfem_mcp_server

Further Documentation

Detailed instructions for installation, port configuration, and integration into various MCP clients can be found in the technical documentation:

Author

Mr. Faulstich is responsible for the quality assurance of the RFEM program and provides customer support.

Links
;