Catenda AI connects AI agents — Copilot, Claude, Gemini and others — to construction project data through an open standard, so teams can ask questions about a project and receive sourced answers without switching AI platforms.
What the MCP server does
Catenda exposes your project context through the Model Context Protocol (MCP) directly, so your own agents — Claude, an in-house tool, or whatever you adopt next — can query the same openBIM data with the same permissions, without a bespoke connector and without copying your data elsewhere.
Key benefits
- Speed — comparing a project's contracts, designs, on-site realities and communications becomes dramatically faster.
- Data security — project data stays in the CDE; nothing trains public models.
- User control — the AI acts as the signed-in user, so your project permissions apply, unchanged.
- Open standard — built on MCP rather than proprietary connectors; works with your existing AI tools.
Data it can reach
IFC models, BCF issues, ISO 19650 document states and asset registers. Answers cite their sources so they can be verified.
Typical use cases
- Reporting across project data
- Specification and compliance checking
- Model deviation tracking
- Missing document identification
- Design change coordination
Endpoint
The MCP endpoint is served at https://mcp.catenda.com/mcp and requires
OAuth authentication.
Resources
Browse available reference material at https://mcp.catenda.com/files, including a sample document.