Model Context Protocol (MCP)

What is Model Context Protocol?

MCP is a protocol that enables AI tools and applications to connect with Novu's data and services in a secure, standardized way. It provides a structured method for AI models to:

• Find and retrieve Novu data (subscribers, workflows, notifications, etc.)
• Access specific tools and functionality provided by Novu
• Maintain context about your Novu workspace when working with AI assistants
• Trigger workflows and manage notification preferences

How MCP Works

Novu hosts a remote MCP server that handles requests from AI tools and provides access to Novu data through a secure interface.

Connection URLs:

• US Region: https://mcp.novu.co/ (Streamable HTTP - Recommended)
• EU Region: https://mcp.novu.co/?region=eu (Streamable HTTP - Recommended)
• Legacy SSE: https://mcp.novu.co/sse (also supported in both regions but not recommended)

When an AI tool connects to Novu's MCP server, authentication verifies the user's API key and permissions, then the tool can access relevant Novu data and functionality.

Available Tools

The Novu MCP Server provides 13 tools for interacting with the Novu API:

Core Operations

• get_api_key_status - Check API key status and region configuration
• get_environments - List development and production environments

Workflow Management

• get_workflows - List all workflows with names and identifiers
• get_workflow - Get detailed workflow configuration and payload structure
• trigger_workflow - Execute workflows with custom data for specific subscribers
• create_workflow - Create new workflows with comprehensive configuration including steps, channels, and validation
• update_workflow - Update existing workflows with modified steps, preferences, and configuration
• delete_workflow - Delete workflows by their unique identifier (irreversible action)

Subscriber Management

• find_subscribers - Search subscribers by email, phone, name, or ID

Notification Analytics

• get_notifications - Get notifications with filtering by channels, workflows, dates
• get_notification - Get specific notification details with execution logs

Preference Management

• get_subscriber_preferences - Get subscriber notification preferences for all channels
• update_subscriber_preferences - Update subscriber channel preferences globally or per workflow

Setup

Prerequisites

• Novu API key from dashboard.novu.co
• MCP-compatible AI assistant (Claude Desktop, Claude.ai, Cursor, Windsurf)
• Node.js installed (for Claude Desktop users)

Authentication

1. Get your API key from the Novu Dashboard
2. Identify your region from the dashboard URL:
   • dashboard.novu.co = US region
   • eu.dashboard.novu.co = EU region

Configuration

{
  "mcpServers": {
    "novu": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://mcp.novu.co/",
        "--header",
        "Authorization:Bearer ${NOVU_API_KEY}"
      ],
      "env": {
        "NOVU_API_KEY": "your-novu-api-key-here"
      }
    }
  }
}

{
  "mcpServers": {
    "novu": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://mcp.novu.co/?region=eu",
        "--header",
        "Authorization:Bearer ${NOVU_API_KEY}"
      ],
      "env": {
        "NOVU_API_KEY": "your-novu-api-key-here"
      }
    }
  }
}

Testing Your Setup

After configuration, test with these commands:

Basic Operations

"Check my Novu API key status"
"Show me all my notification workflows"  
"Find subscriber with email user@example.com"

Troubleshooting

Common Issues:

• Authentication errors: Verify API key is valid and has no extra spaces
• Empty results: Check you're using the correct region (US/EU)
• Connection issues: Restart your AI assistant after configuration changes

Test connectivity:

npx @modelcontextprotocol/inspector

Connect with Transport Type: Streamable HTTP, URL: https://mcp.novu.co/, Headers: Authorization: Bearer your-api-key

For additional help, check the Novu documentation or join our Discord community.