Remote MCP Server on Cloudflare (Without Auth)

This example allows you to deploy a remote MCP server that doesn't require authentication on Cloudflare Workers.

Overview: What is a Remote MCP Server?

A Remote MCP Server provides an interface for AI applications, such as Claude Desktop, Continue, and Cursor, to interact with various tools and data sources through the Model Context Protocol. This protocol standardizes interactions between the AI application and external contexts like databases or APIs, ensuring seamless integration and flexibility.

🔧 Core Features & MCP Capabilities

The Remote MCP Server offers several key features that enhance its compatibility and utility across different AI applications:

Tool Integration: You can define tools using the init() method in src/index.ts via this.server.tool(...).
No Authentication Required: The server is designed to be publicly accessible, making it easier for developers to test and deploy without the complications of authentication.
Integration with Cloudflare Workers: Deploying on Cloudflare Workers ensures fast performance and reliability.

MCP Protocol Flow

The Model Context Protocol operates by establishing a connection between the AI application (MCP client) and the remote server, through which commands are issued and responses are received. Here is how it works:

graph TD
    A[AI Application] -->|MCP Client| B[MCP Protocol]
    B --> C[MCP Server]
    C --> D[Data Source/Tool]
    style A fill:#e1f5fe
    style C fill:#f3e5f5
    style D fill:#e8f5e8

⚙️ MCP Architecture & Protocol Implementation

The architecture of the Remote MCP Server is built around Cloudflare Workers, which provides a scalable and reliable environment for running the server. The key components are:

Tool Initialization: In src/index.ts, tools can be added to theserver using the this.server.tool() method.
Event Source (SSE): The server uses Server-Sent Events (SSE) for bidirectional communication, allowing real-time updates and seamless interaction with MCP clients.

🚀 Getting Started with Installation

Deploying With Cloudflare Workers Console:

Follow this link to deploy your server directly from the console:

This will guide you through deploying the server at a URL like: remote-mcp-server-authless.<your-account>.workers.dev/sse.

Local Deployment via Terminal:

Alternatively, you can set up and run the server locally using the command line:

npm create cloudflare@latest -- my-mcp-server --template=cloudflare/ai/demos/remote-mcp-authless

💡 Key Use Cases in AI Workflows

Data Analysis: Connect your client to a database for real-time data analysis.
API Calling: Integrate with external APIs to fetch or push data.

Example: Real-Time Data Analysis Workflow

Suppose you need to perform real-time data analysis on stock market data using an API endpoint that provides live updates. Here's how you could set up the server:

Define a tool in src/index.ts:

this.server.tool({
  name: "stockMarket",
  description: "Fetches live stock prices from the API.",
  execute: async (request) => {
    const response = await fetch("https://api.example.com/stockmarket");
    return response.json();
  }
});

Connect to the server in an AI application like Claude Desktop.

Example: API Data Fetching Workflow

Another use case is fetching weather data from a weather API:

Define another tool:

this.server.tool({
  name: "weather",
  description: "Returns current weather conditions.",
  execute: async (request) => {
    const response = await fetch("https://api.example.com/weather");
    return response.json();
  }
});

Use the tools.weather in any connected MCP client.

🔌 Integration with MCP Clients

The Remote MCP Server supports integration with several popular AI applications, such as:

Claude Desktop: Fully compatible and ready for use.
Continue: Compatible but lacks some features due to API differences.
Cursor: Currently only supports tools without full server compatibility.

MCP Client Compatibility Matrix

MCP Client	Resources	Tools	Prompts
Claude Desktop	✅	✅	✅
Continue	✅	✅	✅
Cursor	❌	✅	❌

📊 Performance & Compatibility Matrix

The remote architecture ensures high performance and reliability, leveraging the strengths of Cloudflare Workers. The server's SSE implementation supports real-time data updates, making it ideal for applications requiring immediate feedback.

🛠️ Advanced Configuration & Security

MCP Server Configuration Sample

To configure your MCP server:

{
  "mcpServers": {
    "calculator": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-calculator"],
      "env": {
        "API_KEY": "your-api-key"
      }
    }
  }
}

Security Considerations

While the server is designed for public access, it's essential to implement proper security measures such as rate limiting and IP whitelisting.

❓ Frequently Asked Questions (FAQ)

Q: How do I secure my MCP server?
- A: Implement rate limiting and IP checks to prevent abuse.
Q: Can I use this with Continue?
- A: Yes, but some features may not be fully available due to compatibility issues.
Q: How often is the server updated?
- A: Regular updates are made based on community feedback and security patches.
Q: What tools can I add to the server?
- A: Any tool that can be executed asynchronously through an HTTP request.
Q: Can I run this locally for testing?
- A: Yes, you can use npm create cloudflare@latest -- to set up a local instance.

👨‍💻 Development & Contribution Guidelines

To contribute to the Remote MCP Server project:

Fork the repository on GitHub.
Clone the forked version locally.
Make changes and submit a pull request.

The community is encouraged to suggest features, report issues, and improve documentation.

🌐 MCP Ecosystem & Resources

For more information on the Model Context Protocol and other related projects:

This comprehensive guide aims to provide a robust foundation for developers looking to integrate AI applications with remote MCP servers, ensuring flexibility and scalability in their development processes.

Building a Remote MCP Server on Cloudflare (Without Auth)