The PlantDemand MCP server lets AI assistants like Claude, ChatGPT, Cursor, and Microsoft Copilot query your live asphalt plant data over the open Model Context Protocol. Below are the questions producers, schedulers, and IT teams ask most often before connecting their first AI agent. New to MCP? Start with the MCP overview or jump straight to the five-minute quickstart.
What is the PlantDemand MCP server?
The PlantDemand MCP (Model Context Protocol) server is an official endpoint at https://plantdemand.com/mcp that lets AI assistants — including Claude, ChatGPT, Cursor, and Microsoft Copilot — query your live asphalt plant schedule, orders, and production data using natural language. It is a standards-based replacement for custom integrations.
Which AI clients can connect to it?
Any client that speaks the Model Context Protocol can connect, including Claude Desktop, Claude.ai web (via the Connectors feature), Cursor, Microsoft Copilot Studio, Continue, Cline, and a growing list of others. See the client setup guides for step-by-step instructions for the most common ones.
Is my plant data secure when an AI agent queries it?
Yes. Every MCP request is authenticated with a per-account Server-Api-Key that is scoped to your PlantDemand account, transported over HTTPS, and rate-limited at the server. The AI client only sees the data your account is already permitted to see — there is no broader exposure. You can revoke a key at any time from your account settings. Read the full MCP security model for details.
Do I need a paid PlantDemand plan to use the MCP server?
You need an active PlantDemand account with at least one plant configured. The MCP server is included with all paid plans at no additional cost. There is no separate “AI add-on” — if you can log into PlantDemand, you can connect an AI agent to it.
How do I get an API token to use with the MCP server?
Follow the API token walkthrough. The short version: log into PlantDemand, open Account → API Keys, generate a new Server-Api-Key, copy it once (it is only shown one time), and paste it into your AI client’s MCP server configuration. The whole process takes about three minutes.
What can I actually ask the AI agent to do?
Anything an operations manager would ask about the schedule, in plain English. Common examples: “What is loaded for Monday at Plant 3?”, “Which customers ordered more than 500 tons last week?”, “Show me open capacity for Friday.”, “Compare this month’s production to last year.” See real-world MCP use cases for prompts plant managers use day-to-day.
Can the AI write to my schedule, or only read?
Today the MCP server exposes read tools (order_dates, plus REST-backed operational queries) so AI agents can answer questions and produce reports without changing your data. Write tools — creating orders, moving loads, changing capacities — are on the roadmap and will require an explicit per-tool permission grant on your API key before they activate. Nothing the AI does today can modify your schedule.
Does it work with Power BI, Command Alkon, or Bitumio?
Yes. Power BI users can pull MCP-served data into dashboards using the Power BI integration walkthrough. Customers running Command Alkon or Bitumio can merge that operational data with PlantDemand’s schedule view by following the Command Alkon & Bitumio data integration guide.
What is the difference between MCP and a regular REST API?
A REST API requires a developer to write code that calls specific endpoints with specific parameters. MCP is a higher-level protocol that an AI agent speaks natively — the agent discovers available tools at runtime, decides which to call based on a natural-language question, and chains multiple calls together. For non-developers, MCP turns “talk to your plant data” into a copy-paste configuration step instead of an engineering project. The PlantDemand MCP server still exposes the underlying REST endpoints for traditional integrations.
What are the rate limits and supported regions?
The MCP server runs in the same production environment as the rest of PlantDemand and is available globally over HTTPS. Standard plan rate limits apply (typically several requests per second per key, well above interactive use). If you have an enterprise use case that needs higher throughput — for example, batch dashboards or a multi-plant Copilot rollout — contact PlantDemand support and we can scope a dedicated rate.
How do I troubleshoot connection errors?
Most failures are one of three things: (1) the Server-Api-Key header is missing or contains a stale token — regenerate and re-paste it, (2) the client is pointing at the wrong URL — it should be exactly https://plantdemand.com/mcp, or (3) the JSON body is malformed. Start with the five-minute curl quickstart to confirm the server responds, then layer in your AI client. If curl works but your client does not, the issue is in the client configuration.
How do I revoke an AI agent’s access?
Open Account → API Keys in PlantDemand and delete the key the agent is using. Revocation is immediate — the next request from that client will return 401 Unauthorized. If you need to rotate a key without an outage, generate the new key first, update the client configuration, and then delete the old key.
Next steps
- Five-minute curl quickstart — verify the server before configuring a client.
- Client setup guides — Claude, Cursor, Copilot Studio, and more.
- Real-world MCP use cases — what plant managers actually ask.
- API token walkthrough — generate your Server-Api-Key.
This guide is part of the PlantDemand for plant operations hub for asphalt plant operations, scheduling, and sales management.