Skip to main content
Real-Time LCA ships an MCP (Model Context Protocol) server that exposes the platform’s REST and EPD APIs to AI clients. Once connected, an assistant like Claude, ChatGPT, or Le Chat can manage projects, map materials, run automapping, generate LCA reports, and browse the EPD library — all through natural language, with no copy-paste of API specs.

​
Where to find the setup

Full setup instructions for each supported client live in the Developers section of the docs.

Agent ready — MCP server

Hosted and local connection options, plus per-client setup for Claude, Claude Code, Codex, ChatGPT, Le Chat, and LM Studio.

​
What you can do with it

  • Ask an assistant to create or clone a project from a brief description.
  • Have it map BIM types to EPDs and explain the choices.
  • Generate a dashboard summary or report on demand.
  • Search the EPD library and compare candidates conversationally.

​
Getting access

Each user needs a personal API key (sent as the X-API-Key header) to use the hosted MCP server. Contact the platform team to request one.

​
How your data is handled

When you connect an external client, the rules of that client (for example Claude, Mistral, or ChatGPT) determine how your data is handled — this is outside Real-Time LCA’s responsibility, so review your client’s terms. If you run a local model through LM Studio, your data stays on your machine. The MCP server itself logs anonymized user activity on Azure servers in the EU region. See Transparency & privacy for the full picture across every AI feature.