MCP Server


 

Back to Product Wiki

 


MCP Server

MCP (Model Context Protocol) is an open standard that lets AI assistants connect directly to external systems and data sources through a defined set of tools. An MCP server for iTwin IoT exposes the iTwin IoT platform as a set of callable actions an AI can use on a user's behalf.

In practice, this means an AI connected to the iTwin IoT MCP server can:

Query live and historical data

pull sensor readings, device status, alert states, and time-series observations from monitored assets (bridges, dams, buildings, construction sites, etc.)

Bulk actions

Update sensor names, group names, mass create standardized dashboards and more

Manage infrastructure configuration

create and update nodes, devices, sensors, and calculations that make up a digital twin's monitoring network

Set up and manage alerts

configure threshold, correlation, pattern, vibration, or status alerts that flag anomalies in sensor data
Run analytics and ML jobs

trigger regression, trend, classification, or anomaly-detection models against sensor data

Generate reports and visualizations

build dashboards, graphs, and scheduled reports summarizing asset health- Handle documents and files — upload, organize, and retrieve inspection reports, imagery, and other project files tied to monitored assets

Provide product guidance 

it can read and feed iTwin IoT KnowledgeBase articles directly into the chat with references to our articles and images from those articles.

 

Instead of a person manually navigating the iTwin IoT web portal to check on a dam's tilt sensors or configure a new vibration alert, they can just ask the AI in natural language, and the MCP server translates that request into the correct API calls against the platform.

 

To see a complete list of the tools available to the AI Agent please navigate to https://itwiniotai-mcp.bentley.com/docs

You will have to authenticate via your Bentley IMS account to access this list. 

List of supported AI Agents connectors

Claude