OpenAI MCP integration
Connect ChatGPT and Codex to your iQunet measurements. Ask questions, compare recorded data and prepare reports in your AI workspace.
Your machine data. Ready for a conversation.
Ask what changed.See the measurements.
Compare recorded measurements across time. Inspect the signals, ask questions and prepare a report.
The iQunet connection gives ChatGPT or Codex access to your gateway’s measurements and history through OpenAI MCP.
Start with the question that matters
Compare operating trends
Compare a chosen time period across sensors or against an earlier baseline.
Compare the temperature histories of these two sensors over the last seven days. Plot them on the same time scale.
Measurements with their context intact
The assistant reads the gateway’s OPC UA information model, so the values stay connected to the signals they describe.
- Signal identity
- Device identifiers, measurement names and data types keep the selected channels identifiable.
- Time and history
- Timestamps connect values to the recording period. Retrieve several selected histories together.
- Units and quality
- Quality codes accompany the data, with engineering units when supplied by the measurement source.
Inspect known measurements directly or browse the available signals to choose what to analyse. Calculations, charts and reports use the retrieved samples in your AI workspace.
From your gateway to your workspace
iQunet gateway
Measurements and
local history
Secure MCP Tunnel
Authenticated
read-only access
ChatGPT & Codex
Analysis and reports
in your workspace
Your history stays on the gateway
The monitoring database remains local. Measurements requested through MCP are returned to the connected AI client for analysis.
Your organisation manages access
You manage the tunnel, its restricted runtime key and the private workspace app. The gateway dashboard handles connection setup, status and key replacement.
Connect it to your installation
Prepare the workspace
Create a tunnel in your OpenAI organisation and associate it with the intended ChatGPT workspace.
Connect the gateway
Enter the tunnel ID and restricted runtime key in OpenAI MCP, enable the bridge and save.
Add the private app
Use the configured tunnel in your workspace app and confirm the gateway reports a ready connection.
Read a measurement
Start a new ChatGPT or Codex session and discover the available sensors and read tools.
Permissions, gateway configuration and a first measurement.
You need gateway software with OpenAI MCP, the required OpenAI tunnel and workspace permissions, and outbound HTTPS access.
Read access for engineering analysis
- Available operations
- Discover signals, read current values and retrieve history. Device control, acquisition settings and gateway code execution are outside this connection.
- Workspace access
- Access is shared by the users of the tunnel binding and workspace app. Individual chat users do not receive separate sensor permissions.
Anomaly Detection Service
OpenAI MCP brings your assistant to the measurements. The separate Anomaly Detection Service trains and runs models that compare new recordings with learned machine behaviour.