PlantDemand operates a production MCP server that lets authenticated AI agents work with your plant, order, customer, and material data through the same backend that powers the main application. If you have ever wished you could ask an AI assistant “what was our DOT tonnage last week” or “which crew has the most schedule changes in June” without exporting reports first, the PlantDemand MCP server is the connection point.
What is MCP?
MCP stands for Model Context Protocol. It is an open standard introduced by Anthropic in late 2024 that gives AI agents a consistent way to discover available tools and call them safely. In practical terms, MCP is the connector that lets Claude, Copilot Studio, Cursor, Replit Agent, Kiro, and other AI clients work with external systems like PlantDemand without each one needing a custom integration.
The PlantDemand MCP server is an adapter over our existing REST and GraphQL APIs. It does not create a parallel data path or bypass any existing permission checks. The agent only ever sees data the authenticated user or API key is already allowed to access.
Why this matters for asphalt and aggregate plants
Most plant operations data lives behind dashboards, spreadsheets, and reports. Getting an answer to a question like “show me all DOT-grade Superpave production for August” usually means logging into a system, building a filter, and exporting a report. With the PlantDemand MCP server connected to an AI assistant, that question becomes a sentence — and the assistant retrieves the answer using the same business rules and permissions the application already enforces.
This is the foundation of AI for asphalt plant scheduling. For background on what AI can and cannot do in plant operations today, read our companion guide on AI and asphalt plant scheduling.
Where to start
- Quickstart — Connect an AI client to the PlantDemand MCP server in five minutes with curl examples you can run from any terminal.
- Tool reference — Every tool the MCP server exposes, with input schemas and example requests.
- Security & authentication — How the server uses Server-Api-Key authentication, how plant-level permissions flow through, and what your data governance team needs to know.
- Use cases — Real prompts plant managers and dispatchers use against the PlantDemand MCP server today.
- Client setup guides — Step-by-step instructions for connecting Claude Desktop, Claude Code, Cursor, Replit Agent, Microsoft Copilot Studio, Azure AI Foundry, Kiro, ChatGPT, and any other MCP-capable client.
The endpoint
The PlantDemand MCP server is reachable at:
https://plantdemand.com/mcpEvery request must include a Server-Api-Key header. Generating an API key follows the same process as any other PlantDemand API integration — see SOP 4.1 in your customer documentation, or contact your PlantDemand representative.
How it differs from a normal API integration
A traditional REST integration requires you (or your developer) to write code that calls specific endpoints, handles authentication, parses responses, and exposes the data inside whatever application you are building. With MCP, the AI agent itself discovers what tools are available and calls them on demand. You write a prompt, the agent figures out which tool to call, calls it, and uses the response to answer your question.
This shifts the integration burden from you (writing custom code) to the agent (using the MCP standard). For non-technical plant managers, this means AI assistants can answer real operational questions without anyone writing code first.
Connect your scheduling data to AI today
The PlantDemand MCP server is live and available to all customers. If you already have a PlantDemand account, your IT or operations lead can generate an API key and have an AI assistant connected within minutes. If you are evaluating PlantDemand for the first time, see our asphalt plant scheduling software overview or start a free trial.
This guide is part of the the PlantDemand scheduling platform hub for asphalt plant operations, scheduling, and sales management.