Ot Aiops MCP server
{
"mcpServers": {
"aiops-tools-ot-aiops": {
"command": "uvx",
"args": [
"ot-aiops"
]
}
}
}{
"mcpServers": {
"aiops-tools-ot-aiops": {
"command": "uvx",
"args": [
"ot-aiops"
]
}
}
}{
"mcpServers": {
"aiops-tools-ot-aiops": {
"command": "uvx",
"args": [
"ot-aiops"
]
}
}
}{
"servers": {
"aiops-tools-ot-aiops": {
"command": "uvx",
"args": [
"ot-aiops"
]
}
}
}{
"mcpServers": {
"aiops-tools-ot-aiops": {
"command": "uvx",
"args": [
"ot-aiops"
]
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsoninstall uvx ot-aiopsversion 0.3.0
What it does
Industrial-AIOps is a read-first data tap for factory floors that connects to 14 OT protocols (OPC-UA, Modbus, S7, Mitsubishi MC, MTConnect, MQTT, EtherNet/IP, EtherCAT, PROFINET, SECS/GEM, HART-IP, BACnet/IP, IO-Link, and REST layers) and performs deterministic analysis to identify downtime causes and operational issues. It is designed for operations teams and AI agents operating in industrial environments where evidence-backed answers and air-gapped deployments are required.
- Diagnose why a production line stopped by analyzing correlated data from multiple OT protocols and receiving a cited root cause verdict
- Measure OEE and detect alarm floods across connected equipment without sending data off-site
- Survey and configure an industrial site using scan and sample operations before collecting multi-day production data for analysis
Tools this server exposes
| Tool | Does |
|---|---|
| iaiops_scan_plan | Preview network packets that would be sent during a device survey without contacting any device |
| iaiops_scan_run | Execute a device survey and itemise every packet class sent to discover OT assets |
| iaiops_onboard_draft | Generate a device configuration draft from a completed scan for manual review and merging |
| iaiops_collect_run | Collect production data from configured lines over a bounded duration and report what was read and missed |
| iaiops_tags_export | Export discovered tags for manual role assignment and metadata enrichment |
| iaiops_tags_apply | Apply role assignments and metadata back to the tag store with audit trail |
| iaiops_oee_measure | Calculate Overall Equipment Effectiveness from collected production history over a time range |
| iaiops_investigate_open | List open investigations and unresolved issues from collected data |
| iaiops_diag_rca | Perform root cause analysis on collected production events with evidence citations |
| iaiops_rca_narrate | Rephrase a computed root cause verdict using natural language (optional, LLM-dependent) |
| iaiops_onboard_status | Check which of the six deployment stages the site is on and which command advances it next |
| iaiops_doctor | Validate config, secrets, device reachability, and version without contacting devices |
| iaiops_readiness | List which analysis scenarios can run with the current setup and what gaps need to be closed |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.