APS-Agent (AVEVA MCP Server)
MCP serverCommunity
An academic MCP server that lets an agent create and run AVEVA Process Simulation models. By gsi-lab.
6 starsUpdated 5 months agoGitHub data from 5 Nov 2026
At a glance
- Type
- MCP server
- Works with
- AVEVA Process Simulation
- How it connects
- Drives a running copy of the tool
- Access
- Can modify models
- Tool versions
- Not stated in its documentation
- Operating systems
- Windows
- Needs a AVEVA Process Simulation licence
- Yes
- Transport
- Local (stdio), Remote (HTTP)
- AI clients documented
- Claude Desktop, Cursor
- Licence
- MIT
- Language
- Python
What it can do
- Connects to AVEVA Process Simulation through the Python components of the AVEVA SDK (simcentralconnect) and exposes tools such as connecting and creating a simulation.
- Runs over stdio for desktop clients, or SSE and streamable HTTP on a port you choose.
- Accompanies a peer-reviewed paper on an LLM agent for chemical process simulation by Liang, Groll and Sin (2026).
Limitations
- The logic ships as compiled .pyd modules built for Python 3.13 only, so you can't read or change it, and other Python versions fail to import.
- Needs Windows 10 or 11 with AVEVA Process Simulation and its SDK installed separately.
- The README doesn't list the full tool set or supported AVEVA Process Simulation versions.
Install
pip install -r requirements.txt
# Default: stdio transport (for Claude Desktop/Cursor)
python start_aveva_mcp_server.py{
"mcpServers": {
"aveva": {
"command": "C:/Users/xxxx/AppData/Local/miniconda3/envs/your_py313_env/python.exe",
"args": [
"C:/path/to/distribute/start_aveva_mcp_server.py"
]
}
}
}Source: github.com/gsi-lab/APS-Agent#3-configure-claude-desktop-mcp-client
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.