This Modelica MCP Server exposes two primary MCP tools that you can leverage from any MCP-compatible client: modelica_simulate() and modelica_get_diagram_url(). The modelica_simulate() tool runs a complete Modelica simulation workflow end-to-end, taking a natural language or structured prompt and returning simulation results, including state trajectories and derived insights. The modelica_get_diagram_url() tool creates an interactive Modelica diagram via the integrated ODE diagramming capability, returning a URL you can open to inspect component placement, connections, and annotations. To use these tools, connect your MCP client to the server endpoint, then invoke the appropriate functions with the required modelica descriptions or prompts. The server handles model generation, diagram creation, and autonomous simulation orchestration, returning outputs suitable for visualization, further analysis, or downstream AI workflows.

Prerequisites:

• Node.js (recommended LTS) and npm installed on your system
• Access to the MCP server repository or a built server.js entry point

Installation steps:

1. Ensure Node.js and npm are installed:

   • macOS/Linux: curl -fsSL https://deb.nodesource.com/setup_lts.x | sudo -E bash - && sudo apt-get install -y nodejs
   • Windows: install Node.js LTS from https://nodejs.org/

2. Clone the MCP server repository (or obtain the server bundle): git clone https://github.com/orthogonalpub/modelica_simulation_mcp_server.git cd modelica_simulation_mcp_server

3. Install dependencies (if package.json is present): npm install

4. Configure environment variables (example):

   • PORT=8080
   • MCP_TOKEN=your-token-here You can place these in a .env file or export them in your shell.

5. Start the MCP server: npm start

   or, if using a direct node entry point:

   node server.js

6. Verify the server is running by querying the MCP endpoint or visiting http://localhost:8080/docs (if provided by the server).

7. Connect an MCP client and begin using modelica_simulate() and modelica_get_diagram_url().

Tips and common considerations:

• Authentication: If the server requires a token, set MCP_TOKEN in your environment or configure the client to pass the token with each request.
• Port and host: Adjust PORT and host binding as needed to fit your deployment (e.g., behind a reverse proxy or within a container).
• Diagram generation: modelica_get_diagram_url() returns a URL to an interactive Modelica diagram; ensure your client can render or embed external URLs.
• Model generation and simulation: The server automates model generation, simulation execution, and result extraction. You can extend your prompts to request specific outputs (e.g., plots, data exports, or CSV time-series).
• Security: If deploying publicly, enable token-based authentication and consider TLS termination at a reverse proxy.
• Troubleshooting: Check logs for missing dependencies, port conflicts, or authentication failures; ensure that the server.js entry point exists and is accessible.
• Versioning: Keep the MCP server and client tools aligned with MCP specifications to avoid incompatibilities with newer features.
• Environment variables: Use PORT for the listening port and MCP_TOKEN (or equivalent) for secured access; add any additional environment-specific configs as needed.