dxpert.ai — industrial AI-readiness & UNS agents MCP server
Brings industrial AI readiness diagnostics, UNS architecture design, and specialized plant agents to your local assistant via the dxpert API.
claude mcp add --env DXPERT_API_KEY=<DXPERT_API_KEY> --env DXPERT_API_BASE=<DXPERT_API_BASE> --transport stdio dxpertai-industrial-ai-readiness-uns-agents -- npx -y @dxpert/mcpWhat dxpert.ai — industrial AI-readiness & UNS agents does
This stdio MCP server connects your development environment to the dxpert.ai API. It adds tools for industrial domain guidance, including ask_dxpert for questions about OT/IT data architecture and Unified Namespace layout. You can also run readiness assessments via run_diagnostic or trigger specialized factory agents for shift summaries, root cause investigation, alarm triage, and OEE narration.
The package operates without third-party runtime dependencies. It handles the JSON-RPC stdio protocol directly and forwards requests to the cloud using the X-Api-Key header. Releases publish to npm via GitHub Actions OIDC trusted publishing. A built-in csv_to_bundle tool formats standard spreadsheet exports so agents can inspect them immediately.
Every request requires the DXPERT_API_KEY environment variable. Without this credential, the server returns an authorization error. The agents operate strictly on the bundle passed into the tool call: they do not connect to live plant networks, brokers, or historians. A free tier provides 10 advisory questions per month and 5 trial agent runs before requiring a Pro upgrade.
What it helps with
- evaluate plant IT infrastructure for industrial AI readiness
- design Unified Namespace hierarchies and OT/IT topic structures
- generate plant shift reports directly from spreadsheet exports
- investigate root causes of machinery downtime using incident context
- prioritize and triage high-volume industrial alarm streams
Example prompts for the agent
- Ask dxpert how to structure our Unified Namespace topic hierarchy for packaging lines.
- Convert
shift_log.csvinto a payload usingcsv_to_bundleand generate a shift report. - Run the dxpert diagnostic intake to assess our plant readiness for machine learning.
How to install dxpert.ai — industrial AI-readiness & UNS agents
Platform CLI
claude mcp add --env DXPERT_API_KEY=<DXPERT_API_KEY> --env DXPERT_API_BASE=<DXPERT_API_BASE> --transport stdio dxpertai-industrial-ai-readiness-uns-agents -- npx -y @dxpert/mcpReplace <…> with your own values
Platform CLI
codex mcp add dxpertai-industrial-ai-readiness-uns-agents --env DXPERT_API_KEY=<DXPERT_API_KEY> --env DXPERT_API_BASE=<DXPERT_API_BASE> -- npx -y @dxpert/mcpReplace <…> with your own values
Config file
{
"mcpServers": {
"dxpertai-industrial-ai-readiness-uns-agents": {
"command": "npx",
"args": [
"-y",
"@dxpert/mcp"
],
"env": {
"DXPERT_API_KEY": "<DXPERT_API_KEY>",
"DXPERT_API_BASE": "<DXPERT_API_BASE>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"dxpertai-industrial-ai-readiness-uns-agents": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@dxpert/mcp"
],
"env": {
"DXPERT_API_KEY": "<DXPERT_API_KEY>",
"DXPERT_API_BASE": "<DXPERT_API_BASE>"
}
}
}
}Replace <…> with your own values
Platform CLI
gemini mcp add -e DXPERT_API_KEY=<DXPERT_API_KEY> -e DXPERT_API_BASE=<DXPERT_API_BASE> dxpertai-industrial-ai-readiness-uns-agents npx -y @dxpert/mcpReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about dxpert.ai — industrial AI-readiness & UNS agents
How to install dxpert.ai — industrial AI-readiness & UNS agents in Claude Code?
Run claude mcp add mcp -e DXPERT_API_KEY=<DXPERT_API_KEY> -- npx -y @dxpert/mcp. To specify a custom API endpoint, pass --env DXPERT_API_BASE=<DXPERT_API_BASE>.
What does this MCP server do?
It connects your assistant to the dxpert.ai API, enabling UNS data architecture advice, factory AI-readiness diagnostics, and automated analysis agents.
Do the agents connect directly to plant equipment?
No, the agents reason strictly over the data bundle provided in the tool call. They never connect directly to PLCs, message brokers, or plant historians.
