Remote MCP Server on Cloudflare (Without Authentication)

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

Overview: What is the Remote MCP Server?

The Remote MCP Server enables AI applications like Claude Desktop, Continue, Cursor, and others to connect to specific data sources and tools through a standardized protocol. Deployed as a serverless solution using Cloudflare Workers, this server runs without authentication, making it ideal for early-stage development or testing purposes.

🔧 Core Features & MCP Capabilities

The Remote MCP Server supports the Model Context Protocol (MCP) that facilitates seamless integration between AI applications and external tools and data sources. This implementation includes capabilities such as:

• Authentication-Free Deployment: No need to configure authentication mechanisms, making it quick and easy to set up.
• Custom Tool Integration: Tools can be dynamically added or removed by modifying the src/index.ts file or using the this.server.tool(...) method.
• Live Updates: The server supports real-time updates through Server-Sent Events (SSE), allowing for dynamic data feeds and tool responses.

⚙️ MCP Architecture & Protocol Implementation

The architecture of this Remote MCP Server is constructed to be highly modular, enabling easy integration with various APIs and tools. The core components include:

• Cloudflare Workers: Utilizes Cloudflare's edge computing capabilities to provide low-latency performance.
• MCP Endpoint: A defined endpoint that handles incoming MCP requests from applications like Claude Desktop.
• Tool Integration Layer: Allows for adding, removing, or updating tools in real-time.

MCP Protocol Flow Diagram

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 Client Compatibility Matrix

MCP Client	Resources	Tools	Prompts	Status
Claude Desktop	✅	🗾	✗	Full Support
Continue	🗾	💻	❌	Full Support
Cursor	❌	💼	❌	Limited Support

Note: The 紋 symbol indicates full support where the tool is directly compatible, while a ✗ or ❌ symbol indicates partial or no support.

🚀 Getting Started with Installation

To deploy your Remote MCP Server on Cloudflare Workers:

1. Use the button below to get started:

2. Alternatively, you can use the command line:

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

💡 Key Use Cases in AI Workflows

Use Case 1: Real-Time Data Analysis for Financial Applications

In this scenario, an AI application like Continue can connect to a financial data API hosted on the Remote MCP Server. The server continuously updates with real-time market data, allowing the application to perform live analysis and provide insights to traders.

Technical Implementation: Tools are added using the src/index.ts file. An example configuration might look like:

this.server.tool("finance-api", () => {
    // Implement logic for interacting with financial API
});

Use Case 2: Custom Prompts for Creative Writing

Developers can use Cursor, an AI application for creative writing, to connect to a custom prompt database hosted on the Remote MCP Server. The server dynamically updates prompts based on user feedback and interactions.

Technical Implementation: Tools are added or removed using configuration changes in src/index.ts.

🔌 Integration with MCP Clients

To integrate your Remote MCP Server with an AI application like Claude Desktop, follow these steps:

1. Access the Cloudflare AI Playground: https://playground.ai.cloudflare.com/
2. Enter the URL of your deployed MCP server.
3. Use the tools available in the playground by running sample queries or commands.

Similarly, Continue and Cursor can connect to this remote MCP server following similar steps.

📊 Performance & Compatibility Matrix

PerformanceMetric	Cloudflare Workers Edge Locations
Latency	<100ms
Bandwidth	Up to 500KB/s

Client Compatibility	Claude Desktop	Continue	Cursor
Real-Time Updates	✔	✔	✗
Tool Integration	✔	✔	❌
Authentication	No Auth Needed	N/A	N/A

🛠️ Advanced Configuration & Security

• Security: The server is deployed without authentication, which makes it less secure but ideal for development or testing environments.
• Environment Variables: Use environment variables to configure the API key or other sensitive information.

Example configuration:

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

❓ Frequently Asked Questions (FAQ)

Q1: Can I use this server with other AI clients besides those listed in the compatibility matrix?

A1: Yes, while the current implementation is primarily tested and compatible with Claude Desktop, Continue, and Cursor, you can extend support for additional clients by following similar integration steps.

Q2: How do I handle authentication if I need secure access to my MCP server?

A2: For secure access, consider implementing authentication mechanisms within your codebase or using a middleware service like Cloudflare Access, but this requires modifying the existing setup significantly.

Q3: Are real-time updates supported for all connected tools?

A3: Real-time updates are currently supported for most tools in the compatibility matrix. However, some tools might require additional configurations to ensure correct handling of data streams.

Q4: Can I host my MCP server on other platforms besides Cloudflare Workers?

A4: Yes, this implementation is not tied specifically to Cloudflare Workers and can be hosted on any platform that supports Node.js and the Model Context Protocol. You would need to adapt the deployment steps accordingly.

Q5: How do I ensure data privacy when using MCP servers with public endpoints?

A5: For enhanced security, consider encrypting communication channels (e.g., HTTPS), implementing rate limiting, and ensuring sensitive data is not exposed in your environment variables or logs.

👨‍💻 Development & Contribution Guidelines

Contributions are always welcome! If you have any improvements or need additional features added, please submit a pull request with detailed documentation of your changes.

1. Fork the project.
2. Make new commits of your changes.
3. Submit a pull request.

If you find any bugs, please report them in the issue tracker. Your feedback and suggestions are invaluable!

🌐 MCP Ecosystem & Resources

Explore more about Model Context Protocol at modelcontextprotocol.io or join discussions on the official forums. The community is actively growing and collaborating to enhance the capabilities of this protocol.

This Remote MCP Server is designed to enable seamless integration of tools into AI applications, providing developers with a powerful yet flexible solution for building intelligent workflows.