gnnepcsaft_mcp_server

gnnepcsaft_mcp_server

0

GNNEPCSAFT MCP Server facilitates advanced thermodynamic calculations using Graph Neural Networks to estimate ePC-SAFT parameters, providing predictions for properties like density and vapor pressure. It is suitable for pure components, mixtures, and integrates with LLMs for chemistry applications.

GNNEPCSAFT MCP Server

What is GNNEPCSAFT MCP Server?

GNNEPCSAFT MCP Server is an implementation of the Model Context Protocol (MCP) for GNNePCSAFT tools. It enables seamless communication and context management between large language models (LLMs) and clients for advanced thermodynamic calculations. GNNePCSAFT leverages Graph Neural Networks (GNNs) to estimate ePC-SAFT parameters, allowing property predictions such as density and vapor pressure for any molecule, even without experimental data. FeOS is used for PC-SAFT calculations.


Key Features

  • Estimate ePC-SAFT parameters using GNNs
  • Calculate density, vapor pressure, enthalpy of vaporization, critical points, and others
  • Support for pure components and mixtures
  • Automatic data collection from PubChem for any molecule
  • Designed for thermodynamics-aware LLMs

How to Use

Installation

You need uvx installed.

Starting the Server

uvx --from gnnepcsaft-mcp-server gnnepcsaftmcp

Example: Claude Desktop Configuration

{
  "mcpServers": {
    "gnnepcsaft": {
      "command": "uvx",
      "args": ["--from", "gnnepcsaft-mcp-server", "gnnepcsaftmcp"]
    }
  }
}

Use Cases

  • Predicting thermodynamic properties for new or existing molecules
  • Running property calculations for mixtures in research or industry
  • Integrating with LLMs for chemistry and materials science applications
  • Automating data collection and property estimation in pipelines

FAQ

Q: What do I need to run the server?
A: You need Python, uvx, and the GNNEPCSAFT MCP Server package.

Q: Can I use this for mixtures as well as pure components?
A: Yes, the server supports both pure components and mixtures.

Q: Where does the molecular data come from?
A: The server can automatically fetch molecular information from PubChem.

Q: What calculations are supported?
A: Density, vapor pressure, enthalpy of vaporization, critical points, and ePC-SAFT parameter estimation.

Q: Is this open source?
A: Yes, it is licensed under the GNU General Public License v3.0.


License

GNU General Public License v3.0