Built for engineers who can't afford to miss things
MEPCheck AI is an AI-powered quality control tool for mechanical, electrical, and plumbing engineers. It reviews construction documents for design errors, code violations, and coordination conflicts — in minutes, not days.
Why we built this
MEP design errors are expensive. A missed panel overload, an undersized duct, a missing cleanout — discovered in the field, these issues cost far more to fix than they would have on paper. Yet the QA/QC process that catches them is manual, time-consuming, and inconsistently applied across firms of all sizes.
MEPCheck AI exists to make rigorous QA/QC accessible on every project — not just the large ones with dedicated review teams and extended schedules. By applying the same systematic checklist on every submission, it catches the errors that fatigue and time pressure cause humans to miss.
MEPCheck AI is an AI-assisted tool, not a licensed engineering service. Every finding should be reviewed and confirmed by the engineer of record before any design decision is made.
What it checks
Every review covers all four MEP disciplines simultaneously, applying the following default codes unless the reviewer specifies otherwise:
- Mechanical — duct sizing vs. design CFM, equipment schedule consistency, supply/exhaust balance, maintenance clearances. Default code: ASHRAE 90.1-2022, IMC 2021.
- Electrical — panel loading vs. 80% continuous rule, wire sizing vs. breaker amperage, lighting power density, arc flash labeling, missing disconnects. Default code: NEC 2023.
- Plumbing — pipe sizing for fixture units, venting configurations, cleanout locations, backflow prevention, trap primers. Default code: IPC 2021.
- Coordination — cross-discipline conflicts in ceiling and wall cavities, structural penetrations, equipment access routes, RCP conflicts.
Custom code editions and jurisdictional amendments can be specified in the Reviewer Notes field before each review.
How it works
You upload your drawings, schedules, and specifications directly in the browser. The files are sent to an AI model that reads them the way a senior engineer would — checking for the same issues, applying the same codes, producing a structured report with findings tagged by discipline, severity, sheet location, and code reference.
Nothing is stored. Once your report is generated, the uploaded files are discarded. Your documents never feed into AI training or get shared with third parties.
Our values
Contact
For questions, enterprise inquiries, or feedback, visit our contact page .