Multi-tenant MCP Server

Integration App MCP Server (SSE)

This is a remote implementation of the MCP (Model Context Protocol) server that exposes tools powered by Integration App. It allows clients to connect and access tools from active connections, using the MCP SSE transport.

To implement your own MCP client, see our example AI Chat Agent:

• AI Chat Agent (MCP Client application)

Prerequisites

• Node.js (v14 or higher)
• npm or yarn
• An Integration App account with a valid JWT token

Installation

1. Clone the repository:
   git clone https://github.com/integration-app/mcpservice cd mcpservice
2. Install dependencies:
   npm install
3. Build the project:
   npm run build

Configuration

Local Development

To run the server locally, start it with:

The server will run on http://localhost:3000.

Deployment

To deploy the server to a production environment (e.g., Heroku), follow these steps:

1. Ensure your environment variables are set:
   • PORT: The port on which the server will run (default: 3000)
   • NODE_ENV: Set to production for production environments
2. Deploy your application using your preferred hosting service (e.g., Heroku, AWS, etc.).
3. Once deployed, your server will be accessible at a URL like:
   https://your-app-name.herokuapp.com/

Connection URL

To connect to the MCP server, use the following URL format:

Or, if the server is running locally:

Cursor Configuration

To use this server with Cursor, update your ~/.cursor/mcp.json file:

Restart Cursor for the changes to take effect.

Claude Configuration

Anthropic only allows SSE MCP tranports to Claude with MAX plan or higher. To use this server with Claude, update your ~/claude_desktop_config.json file:

MCP Information

• The server fetches tools from all active connections associated with the provided token.
• The server supports both SSE (Server-Sent Events) and Streamable HTTP transports.

Troubleshooting

• Ensure your JWT token is valid and has the necessary permissions.
• Check server logs for any errors or issues during startup or connection attempts.
• Verify that your deployment environment has the correct environment variables set.