Remote MCP Server on Cloudflare (Without Auth)

This example demonstrates how to deploy a remote Model Context Protocol (MCP) server without authentication on Cloudflare Workers, allowing AI applications to connect seamlessly using the standard protocol.

Overview: What is a Remote MCP Server?

An MCP Server enables AI applications like Claude Desktop, Continue, Cursor, and others to interact with specific data sources and tools via a standardized protocol. Deploying this remote MCP server ensures that these applications can leverage external services or local tools by adhering strictly to the Model Context Protocol.

🔧 Core Features & MCP Capabilities

The Remote MCP Server enables AI clients to connect without requiring authentication. This setup is particularly useful for shared environments or testing scenarios where user authentication may not be necessary. Key features include:

• Deployability: Easy deployment via Cloudflare Workers with a simple button click, integrating seamlessly into the Cloudflare ecosystem.
• Customizability: Tools can be added through custom methods in src/index.ts, allowing developers to integrate their preferred tools into this MCP server infrastructure.

⚙️ MCP Architecture & Protocol Implementation

The architecture of this remote MCP server ensures compatibility with various AI applications. The Model Context Protocol (MCP) flow involves an AI Application interacting with a MCP Client, which then communicates via the MCP Server to access Data Source/Tools. Here's a diagram to illustrate:

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

Real-World AI Workflow Use Case 1: Financial Analysis

Imagine a scenario where an investment firm uses the remote MCP server to connect with financial APIs. The firm can use tools within the server to fetch real-time market data, analyze trends, and generate reports directly from Claude Desktop or another AI client.

Real-World AI Workflow Use Case 2: Environmental Monitoring

In environmental monitoring applications, a remote MCP server could integrate with weather sensors and machine learning models to predict climate patterns. This setup allows users like researchers to quickly access relevant tools without extensive manual setup, enhancing the efficiency of their workflows.

🚀 Getting Started with Installation

Deploying your MCP server can be done through two methods: using the Cloudflare Worker deployment button or manually via the command line.

Using the Cloudflare Worker Deployment Button

1. Click on the Deploy to Workers button.
2. Your server will be deployed at a URL like remote-mcp-server-authless.<your-account>.workers.dev/sse.

Manual Command Line Installation

Run the following command to set up your MCP server locally:

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

💡 Key Use Cases in AI Workflows

This Remote MCP Server is ideal for various AI application use cases, including:

• Data Aggregation and Analysis: Integrate with external APIs for seamless data fetching.
• Tool Integration: Customize the server to include specific tools required by your AI applications.
• Automated Reporting: Generate reports based on tool execution results without manual intervention.

🔌 Integration with MCP Clients

To use this Remote MCP Server, follow these steps:

1. Go to https://playground.ai.cloudflare.com/
2. Enter the URL of your deployed MCP server (remote-mcp-server-authless.<your-account>.workers.dev/sse).
3. Explore and use the tools integrated into the server directly from the playground!

Similarly, integrate this server with Claude Desktop by configuring it as follows:

{
  "mcpServers": {
    "calculator": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "http://localhost:8787/sse" // or remote-mcp-server-authless.your-account.workers.dev/sse
      ]
    }
  }
}

📊 Performance & Compatibility Matrix

The compatibility matrix for this Remote MCP Server outlines its support across different MCP clients:

MCP Client	Resources	Tools	Prompts	Status
Claude Desktop	✅	✅	✅	Full Support
Continue	✅	✅	✅	Full Support
Cursor	❌	✅	❌	Tools Only

🛠️ Advanced Configuration & Security

For advanced configurations, consider the following:

• Environment Variables: Securely store API keys and other secrets using environment variables.
• Customization: Extend the server's functionality by adding more tools in src/index.ts.

FAQ

1. Can I connect this server to multiple AI clients simultaneously?

Yes, but ensure that your server has appropriate concurrency support.

2. Why might my tool not appear in the MCP client?

Check that you've correctly added and configured the tool within the init() method of src/index.ts.

3. How do I secure this server against unauthorized access?

Utilize Cloudflare's built-in security features for worker functions to ensure only expected clients can access your server.

4. Is there a limit on the number of tools I can add?

The platform imposes no explicit limits, but practical limitations may arise based on tool complexity and resource usage.

5. Can this server be deployed in other cloud environments besides Cloudflare?

Yes, it can be adapted for deployment on platforms like AWS or GCP by adjusting the hosting environment config.

👨‍💻 Development & Contribution Guidelines

Contributions to this project are encouraged. To contribute:

1. Fork the repository and clone it locally.
2. Make changes and push your changes to a branch.
3. Open a pull request for review and merge.

🌐 MCP Ecosystem & Resources

Explore more resources and tools in the broader Model Context Protocol ecosystem, including official documentation and community forums.

• Cloudflare AI Documentation
• MCP Client Compatibility List

By leveraging this Remote MCP Server, developers can enhance their AI applications with robust tool integrations and seamless protocol compatibility.