fluent-research-mcp

MCP serverCommunity

A local MCP server that runs Fluent cases through PyFluent or batch journals and keeps an audit trail of every run. By the-syz.

6 starsUpdated 5 months agoGitHub data from 5 Nov 2026

At a glance

Type: MCP server
Works with: Ansys Fluent / CFX
How it connects: Drives a running copy of the tool
Access: Can modify models
Tool versions: Smoke-tested on Fluent 2025 R2 (3ddp, serial batch); other releases not stated
Operating systems: Windows
Needs a Ansys Fluent / CFX licence: Yes
Transport: Local (stdio)
AI clients documented: Codex
Licence: MIT
Language: Python

What it can do

Checks the environment first, including Python, PyFluent, the Fluent executable, licence hints and a C compiler.
Runs existing case and data files through either a PyFluent backend or a Fluent batch journal backend.
Saves an input snapshot, journal, logs, status file, summary and report for each run so failures can be traced.
Writes UDF source files as a reviewable diff and compiles and loads them only after review.
Exports a results inventory, CSV data and optional images.

Limitations

README is in Chinese, and the only documented client configuration is a Codex TOML entry.
Install instructions use Windows PowerShell; other platforms are not described.
Needs Python 3.11 or later and a licensed Fluent install for any solver work.

Install

python -m venv .venv
.\.venv\Scripts\Activate.ps1
python -m pip install -U pip
python -m pip install -e .

Source: github.com/the-syz/fluent-mcp

[mcp_servers.fluent-research-mcp]
command = '<repo>\\.venv\\Scripts\\python.exe'
args = ['-m', 'fluent_research_mcp.server']
startup_timeout_sec = 30
tool_timeout_sec = 600

Source: github.com/the-syz/fluent-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.