OpenStudio-MCP: A Model Context Protocol (MCP) Server for AI Agent-Driven Building Energy Modeling with the OpenStudio SDK: Article No. 103020

Brian Ball; Nicholas Long; Katherine Fleming; David Goldwasser

doi:10.1016/j.softx.2026.103020

OpenStudio-MCP: A Model Context Protocol (MCP) Server for AI Agent-Driven Building Energy Modeling with the OpenStudio SDK: Article No. 103020

Building Technologies and Science Center

Research output: Contribution to journal › Article › peer-review

Abstract

Building energy modeling (BEM) can test design options before construction or renovation, but effective OpenStudio(R) use often requires programming Measures and deep modeling expertise. OpenStudio-MCP is an open-source Model Context Protocol (MCP) server that gives AI agents validated access to the OpenStudio SDK through 197 tools and a runtime knowledge layer. Unlike BEM servers focused on lower-level edits to existing models, it supports model creation, HVAC synthesis, Measure reuse and authoring, EnergyPlus simulation, and results analysis from natural language. It also provides sandboxed multi-user deployment, validation, compact result extraction, and a reproducible, artifact-graded benchmark across six large language models from two vendors.

Original language	American English
Number of pages	9
Journal	SoftwareX
Volume	36
DOIs	https://doi.org/10.1016/j.softx.2026.103020
State	Published - 2026

NLR Publication Number

NLR/JA-5500-101191

Keywords

agent skills
building energy modeling
model context protocol
OpenStudio

Access to Document

10.1016/j.softx.2026.103020

Cite this

@article{5153ea48f10946c3a4ebbc2dfe89f4d9,

title = "OpenStudio-MCP: A Model Context Protocol (MCP) Server for AI Agent-Driven Building Energy Modeling with the OpenStudio SDK: Article No. 103020",

abstract = "Building energy modeling (BEM) can test design options before construction or renovation, but effective OpenStudio(R) use often requires programming Measures and deep modeling expertise. OpenStudio-MCP is an open-source Model Context Protocol (MCP) server that gives AI agents validated access to the OpenStudio SDK through 197 tools and a runtime knowledge layer. Unlike BEM servers focused on lower-level edits to existing models, it supports model creation, HVAC synthesis, Measure reuse and authoring, EnergyPlus simulation, and results analysis from natural language. It also provides sandboxed multi-user deployment, validation, compact result extraction, and a reproducible, artifact-graded benchmark across six large language models from two vendors.",

keywords = "agent skills, building energy modeling, model context protocol, OpenStudio",

author = "Brian Ball and Nicholas Long and Katherine Fleming and David Goldwasser",

year = "2026",

doi = "10.1016/j.softx.2026.103020",

language = "American English",

volume = "36",

journal = "SoftwareX",

issn = "2352-7110",

publisher = "Elsevier B.V.",

}

TY - JOUR

T1 - OpenStudio-MCP: A Model Context Protocol (MCP) Server for AI Agent-Driven Building Energy Modeling with the OpenStudio SDK

T2 - Article No. 103020

AU - Ball, Brian

AU - Long, Nicholas

AU - Fleming, Katherine

AU - Goldwasser, David

PY - 2026

Y1 - 2026

N2 - Building energy modeling (BEM) can test design options before construction or renovation, but effective OpenStudio(R) use often requires programming Measures and deep modeling expertise. OpenStudio-MCP is an open-source Model Context Protocol (MCP) server that gives AI agents validated access to the OpenStudio SDK through 197 tools and a runtime knowledge layer. Unlike BEM servers focused on lower-level edits to existing models, it supports model creation, HVAC synthesis, Measure reuse and authoring, EnergyPlus simulation, and results analysis from natural language. It also provides sandboxed multi-user deployment, validation, compact result extraction, and a reproducible, artifact-graded benchmark across six large language models from two vendors.

AB - Building energy modeling (BEM) can test design options before construction or renovation, but effective OpenStudio(R) use often requires programming Measures and deep modeling expertise. OpenStudio-MCP is an open-source Model Context Protocol (MCP) server that gives AI agents validated access to the OpenStudio SDK through 197 tools and a runtime knowledge layer. Unlike BEM servers focused on lower-level edits to existing models, it supports model creation, HVAC synthesis, Measure reuse and authoring, EnergyPlus simulation, and results analysis from natural language. It also provides sandboxed multi-user deployment, validation, compact result extraction, and a reproducible, artifact-graded benchmark across six large language models from two vendors.

KW - agent skills

KW - building energy modeling

KW - model context protocol

KW - OpenStudio

U2 - 10.1016/j.softx.2026.103020

DO - 10.1016/j.softx.2026.103020

M3 - Article

SN - 2352-7110

VL - 36

JO - SoftwareX

JF - SoftwareX

ER -