My MCP Server is an MCP server that provides a remote, serverless implementation of the Model Context Protocol running on Cloudflare Workers without authentication requirements. It connects AI clients directly to custom tools hosted at Cloudflare edge locations via Server-Sent Events endpoints. Developers building lightweight AI agent workflows or prototyping tools use this server to deploy endpoints without managing persistent server infrastructure or implementing authentication middleware. The server enables users to register custom functions using Cloudflare Workers and execute them from remote clients like the Cloudflare AI Playground or local desktop clients. By providing a deployable template, it simplifies the process of exposing backend functions to large language models. Custom tools are defined directly within the Worker codebase, making it straightforward to add calculator utilities, web request handlers, or edge computation functions that AI assistants can invoke directly during conversational tasks.
Category: Other & General Purpose
Tags: cloudflare, deployment, serverless, workers
bash npm create cloudflare@latest -- my-mcp-server --template=cloudflare/ai/demos/remote-mcp-authless 2. Define your tools inside the init() method of src/index.ts using this.server.tool(...). 3. Deploy the project to Cloudflare Workers to obtain your remote Server-Sent Events URL, such as remote-mcp-server-authless.<your-account>.workers.dev/sse. 4. To connect to Claude Desktop, open Claude Desktop and navigate to Settings > Developer > Edit Config. 5. Add the server entry using mcp-remote under mcpServers in your configuration file: json { "mcpServers": { "calculator": { "command": "npx", "args": [ "mcp-remote", "https://remote-mcp-server-authless.<your-account>.workers.dev/sse" ] } } } 6. Restart Claude Desktop to access your tools.Part of MCP Servers
My MCP Server is an implementation of a Model Context Protocol server designed to run serverlessly on Cloudflare Workers. It uses Server-Sent Events to provide communication between AI assistants and edge tools without requiring authentication.
You can scaffold the project locally by running npm create cloudflare with the cloudflare ai demos remote-mcp-authless template. Then customize your tools in the index file and deploy it using Cloudflare Workers commands to obtain a remote URL.
It connects natively to remote clients such as the Cloudflare AI Playground. It can also connect to local desktop clients like Claude Desktop by running the mcp-remote proxy package configured within the client settings.
It serves edge-based tools to AI clients using the Model Context Protocol. You can register custom tools inside the server code, such as computational functions, APIs, or data parsers, allowing the AI to call them dynamically.