OpenFOAM MCP (milsonson)

MCP serverCommunity

A template-driven OpenFOAM server that goes from a prompt to case, preflight check, solve and residual plot. By milsonson.

2 starsUpdated 8 months agoGitHub data from 5 Nov 2026

At a glance

Type: MCP server
Works with: OpenFOAM
How it connects: Drives a running copy of the tool
Access: Can modify models
Tool versions: unknown (needs a sourced OpenFOAM environment; mpirun for parallel runs)
Operating systems: Not stated in its documentation
Needs a OpenFOAM licence: No
Transport: Local (stdio), Remote (HTTP)
AI clients documented: Not stated in its documentation
Licence: Not stated in its documentation
Language: Python

What it can do

Analyses a plain-language problem, recommends a template and fills missing parameters.
Creates standard 0/, constant/ and system/ case folders from templates covering incompressible, heat-transfer, multiphase and compressible cases.
Runs a preflight and stability check, solves in serial or parallel, and falls back to serial if a parallel run fails.
Plots residuals with matplotlib, and offers health, job-status and artefact endpoints over HTTP.

Limitations

Mesh generation covers blockMesh only; the snappyHexMesh interface is reserved but not built.
README is in Chinese, the repository has no licence file, and the default transport is SSE.

Install

cd /path/to/openfoam-mcp
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
pip install -e .

Source: github.com/milsonson/openfoam-mcp

{
  "mcpServers": {
    "openfoam": {
      "command": "/home/your_user/openfoam-mcp/.venv/bin/python",
      "args": [
        "/home/your_user/openfoam-mcp/run_server.py"
      ]
    }
  }
}

Source: github.com/milsonson/openfoam-mcp

MCP servers and skills run code on your computer. Review the source before installing.

AI-generated models, drawings and results must be checked by a qualified engineer before use.