monday.com offers two distinct MCP (Model Context Protocol) servers and a set of native integrations with popular AI tools. This page explains what each one does, when to use it, and how they work together.

The MCP landscape

Platform MCP

The Platform MCP is an open-source MCP server maintained by the monday.com AI team. It lets AI agents read and write monday.com data through the GraphQL API — creating items, updating columns, querying boards, managing users, and more.

Common use cases:

• Letting an AI assistant manage tasks, update statuses, and create boards on behalf of a user
• Building custom agents that understand a monday.com account and interact with it programmatically
• Automating reporting, triage, and project management workflows

Configuration

Connect using the hosted MCP, which requires no local setup:

{
  "mcpServers": {
    "monday-mcp": {
      "url": "https://mcp.monday.com/mcp"
    }
  }
}

See the Platform MCP README for the full tool list, authentication options, dynamic API tools (beta), and advanced configuration.

Calling the API through the MCP

One of the most useful capabilities of the Platform MCP is calling the monday.com API directly. You can control the API version by setting headers in your MCP configuration:

{
  "mcpServers": {
    "monday-mcp": {
      "url": "https://mcp.monday.com/mcp",
      "headers": {
        "Api-Version": "2026-07"
      }
    }
  }
}

Using a personal API token is particularly useful when you need to switch between monday.com accounts frequently, since each token is scoped to a specific account.

📘

For development workflows, it is recommended to create a dedicated skill or script that calls the API directly rather than routing every call through the MCP. The MCP is best suited for agent-driven interactions where the AI decides what to call and when.

Apps MCP

The Apps MCP is a separate MCP server purpose-built for the monday apps framework. It does not interact with boards or items — it manages apps, versions, features, deployments, and environment variables.

Common use cases:

• Scaffolding a new monday.com app with features (board views, widgets, integrations)
• Checking deployment status, promoting versions, and managing app storage
• Searching official monday.com apps documentation from inside an AI assistant

Configuration

{
  "mcpServers": {
    "monday-apps-mcp": {
      "command": "npx",
      "args": [
        "@mondaydotcomorg/monday-api-mcp",
        "-t", "your_monday_api_token",
        "--mode", "apps"
      ]
    }
  }
}

📘

The --mode apps flag is what switches the server from Platform MCP mode to Apps MCP mode.

Using both MCPs together

Most developers building monday.com apps connect both the Platform MCP and the Apps MCP to their IDE (Cursor, Claude Code, etc.). With both active, an AI agent can:

1. Read board schemas and item data to understand the account context (Platform MCP)
2. Scaffold, deploy, and promote app versions (Apps MCP)
3. Set environment variables and check deployment logs (Apps MCP)
4. Test the app's behavior by querying live data (Platform MCP)

This combination lets an agent automate a large portion of the app-building workflow end to end.

Native integrations

monday.com provides pre-built integrations with several popular AI tools. These are the simplest way for end users to connect monday to an AI assistant.

MCP resources

• MCP home page
• Get started with monday MCP
• Ready-to-use monday MCP prompts
• monday MCP available tools

Common questions

How do I connect [tool X] to the monday MCP? If your tool has a native integration listed above, follow the corresponding guide. Otherwise, use the Platform MCP — either the hosted URL or the local npx setup — and add it to your tool's MCP configuration.

What's the difference between the Platform MCP and the Apps MCP? The Platform MCP lets agents interact with monday.com data (boards, items, users). The Apps MCP lets agents manage monday.com apps (create features, deploy code, promote versions). They serve different purposes and can run side by side.

Do I need both MCPs? Only if you're building monday.com apps. If you just want an AI assistant to manage your work in monday, the Platform MCP (or a native integration) is all you need.