Description
Model Context Protocol (MCP) is transforming how AI assistants interact with engineering data. Instead of building separate apps, you expose your data as MCP tools and any large language model (LLM) connects natively. In this session, we'll show how WSP built a production MCP server using FastMCP and Python that exposes AEC Data Model GraphQL APIs as intelligent tools enabling engineers to query live Autodesk Forma Revit models conversationally via the Claude Desktop app or GPT-4. Ask "What is the total embodied carbon for Level 3?" or "List all structural columns over 5m" and you'll get live answers directly from Autodesk Forma (formerly Autodesk Construction Cloud), with no separate UI. This was built and demonstrated at the Autodesk APS Accelerator Bangalore 2026. Attendees will leave with a replicable MCP server blueprint for Autodesk Forma.
Key Learnings
- Learn how to build a production MCP server using FastMCP that exposes AEC Data Model GraphQL APIs as callable AI tools.
- Learn how to connect the Claude Desktop app or GPT-4 to live Autodesk Forma Revit models using MCP protocol and Autodesk Platform Services OAuth 2.0.
- Learn how to design MCP tool schemas for BIM queries including QTO, embodied carbon, compliance, and cost estimation.
- Learn how to deploy and test an MCP server against a real WSP Autodesk Forma project with live GraphQL queries and AI responses.
Downloads
Tags
Product | |
Industries | |
Topics |
People who like this class also liked
