Build faster with AI using the new monday apps MCP

AI coding agents and assistants like Cursor, Claude, Copilot, and Gemini Code are already part of how developers build apps. But until now, they couldn’t interact with the monday apps framework itself - you still had to jump back into the Developer Center to create features, manage versions, or deploy updates.

The monday apps MCP (Model Context Protocol) changes that by connecting your AI agent directly to the monday apps framework, giving it secure access to build, configure, and deploy monday apps faster than ever - all without leaving your development environment.

What it means for you

The monday apps MCP gives AI agents deep access to the monday apps framework - exposing the same tools you use to create apps, add features, manage versions, and handle deployments.
This means you can scaffold new apps in minutes, iterate on features instantly, and manage your entire app lifecycle through natural language conversations inside your coding agent.

Through the MCP, your AI agent can:

• Create and configure apps - Name, describe, and add features like board views, dashboard widgets, or custom columns.
• Manage versions and promotions - List all versions, see which one’s live, and promote drafts to production.
• Handle monday code deployments - Check deployment status, view logs, and manage environment variables.
• Search and export app storage - Query data by keyword and export results as JSON or CSV for analysis or migration.

How it helps you build better

Instead of clicking through multiple screens to set up features, deploy code, or check logs, you can describe what you want - and your AI agent handles the rest.

You can ask:

• "Create a new monday app called 'Customer Health Analyzer'"
• "Add a board view and a dashboard widget to this app"
• "Check the deployment status and show me any errors"
• "Show me all versions of app 12345 and promote the latest draft if the deployment status is successful"
• Your AI agent executes it directly through the apps framework - no tab-switching, no setup friction.
• That means more time for design, strategy, and customer value, and less time on boilerplate configuration.

Your AI agent executes it directly through the apps framework - no tab-switching, no setup friction.
That means more time for design, strategy, and customer value, and less time on boilerplate configuration.
 

How to get started

1. Connect your AI coding agent
   Compatible with Cursor, Claude Code, VS Code, Copilot, Gemini Code, and other MCP-supported agents.
   Requirements: Node.js v20+, npm v5.2+, and a monday.com API token.
   Setup: Follow the integration guide for detailed instructions.
    
2. Configure your MCP client to use monday-api-mcp with --mode apps. Example configuration (format may vary by client):

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

   Replace your_monday_API_token with your actual monday.com API token. The MCP uses your existing monday.com API token and permissions.
   Your AI agent can only perform actions you’re authorized to take within the apps framework - nothing more.
    

3. Create your first AI-assisted app
   Prompt your AI agent to create a test app, add a dashboard widget, or manage your existing app versions - all in one workflow.

Go even further with your AI agent

Once you’ve got the basics running, you can take it a step further. By giving your AI agent access to additional monday developer tools and resources, it can learn from our guidelines, use real examples, and generate code and design that are aligned with monday.com best practices.

Add these tools to your mcpServers configuration to unlock the fullstack:

• monday-welcome-apps - Gives your AI access to simple example apps to understand basic structure.
• monday-kitchen-sink - A comprehensive showcase of features and capabilities for the AI to reference.
• monday-code-quickstarts - Boilerplate code for different languages and frameworks.
• monday-code-python-sdk - Context on using our official Python SDK for backend logic.
• monday official design system - Connects your AI to the Vibe Design System, ensuring generated UIs look native to monday.com.
• monday-platform-mcp - Gives your AI access to read/write data on your account (create test boards, items, etc) to verify your app works.

Suggested fullstack configuration:

{
  "mcpServers": {
    "monday-apps-mcp": {
      "command": "npx",
      "args": ["@mondaydotcomorg/monday-api-mcp", "-t", "YOUR_TOKEN", "--mode", "apps"]
    },
    "monday-platform-mcp": {
      "command": "npx",
      "args": ["@mondaydotcomorg/monday-api-mcp", "-t", "YOUR_TOKEN"]
    },
    "monday-welcome-apps": { 
      "url": "https://gitmcp.io/mondaycom/welcome-apps" 
    },
    "monday-kitchen-sink": { 
      "url": "https://gitmcp.io/mondaycom/monday-kitchen-sink" 
    },
    "monday-code-quickstarts": { 
      "url": "https://gitmcp.io/mondaycom/monday-code-quickstarts" 
    },
    "monday-code-python-sdk": { 
      "url": "https://gitmcp.io/mondaycom/monday-code-python-sdk" 
    },
    "vibe": { 
      "command": "npx", 
      "args": ["@vibe/mcp"] 
    }
  }
}

Try it and share your experience

We’re excited to see what you’ll build when AI becomes part of your monday app development flow.

Share with us your experiences - from apps you’ve scaffolded using AI to ideas for new MCP tools or workflows that could make building even smoother.