CoolProp MCP servers and AI agent skills

How AI connects to CoolProp

An MCP server or skill can only do what CoolProp lets programs do. CoolProp exposes 3 automation interfaces:

• High-level interface (PropsSI) (Python): PropsSI returns one property from two state inputs and a fluid name, for example a saturation temperature from pressure and quality. It covers pure fluids, pseudo-pure fluids and mixtures, lets you force a phase, and can hand the call to REFPROP by prefixing the fluid name when REFPROP is installed. Input pairs differ in speed: temperature and density are quickest, while pressure and enthalpy are much slower, which matters if an agent loops over thousands of points. Documentation
• Low-level AbstractState interface (Python): For repeated calls the low-level AbstractState interface skips string parsing and is faster, and it gives direct access to partial derivatives. Neither listing uses it yet; it's the route to take if you extend the MCP server for cycle calculations. Documentation
• Wrappers for other languages and spreadsheets (Other): CoolProp 8.0.0 ships wrappers for MATLAB, Octave, C#, Java, R, Julia, Modelica, LabVIEW, Mathcad, EES, Excel and LibreOffice, among others, plus static and shared libraries. An agent working in Excel or MATLAB can call the same property functions without Python. Documentation

What Open source (CoolProp developers) offers

CoolProp is maintained as an open-source project, and its documentation for version 8.0.0 covers the property interfaces and wrappers but nothing on AI assistants, MCP or agent skills. Both listings here are independent community work built on the public library.

Checked 4 Nov 2026. Sources: CoolProp wrappers documentation (8.0.0)

What you can do with CoolProp and AI

• Get the full state of a refrigerant at a pressure and temperature, including enthalpy, entropy, density, transport properties and phase.
• Find saturated liquid and vapour states, latent heat and temperature glide for a blend such as R407C at a given pressure.
• Compare fluids on critical point, normal boiling point, GWP and ODP before choosing a refrigerant.
• Work out moist-air enthalpy, humidity ratio, dew point and wet-bulb temperature for an HVAC calculation.
• Write a Python script that tabulates properties along a process path with PropsSI, following the skill's examples.

Taken from the listings' own documentation. Check each listing for exactly what it supports.

Installing

The listings here document setup for Claude Code and Claude Desktop. Install steps, copied from each project's own documentation, are on each listing's page. Most CoolProp integrations run on your own computer next to CoolProp, so you need CoolProp installed and licensed on the same machine unless the listing says otherwise.