- Home
- MCP servers
- Remote MCP Server Authless
Remote MCP Server Authless
- typescript
0
GitHub Stars
typescript
Language
7 months ago
First Indexed
3 months ago
Catalog Refreshed
Documentation & install
Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.
You can deploy and run a remote MCP server that you connect to from compatible MCP clients. This server exposes its tools via a remote endpoint over HTTP/SSE, so you can use it from playgrounds, IDEs, or desktop clients without embedding tools directly in your local environment.
How to use
Connect to your remote MCP server from an MCP client by pointing the client at the server URL provided after deployment. You can then invoke tools remotely as if they were local, and results stream back to you through the standard MCP interface. In practice, you deploy the server, then configure your client or playground to access the server’s SSE endpoint and use the available tools as needed.
How to install
Prerequisites: you need Node.js installed on your machine to run the deployment command.
-
Ensure Node.js is installed on your system. You can verify with
node -vandnpm -v. -
Run the deployment command to create the remote MCP server using the Cloudflare template.
npm create cloudflare@latest -- my-mcp-server --template=cloudflare/ai/demos/remote-mcp-authless
Additional sections
Server connection URL. After deployment, you will access the MCP server at a remote URL of the form remote-mcp-server-authless.<your-account>.workers.dev/sse. Use this URL in MCP clients or in the Cloudflare AI Playground to connect and start using your tools remotely.
Connecting from a local client. To connect a local client to your remote MCP server, you typically configure the client to target the server’s SSE endpoint. For example, a client could be configured to delegate tool calls through the remote server URL, allowing you to execute tools remotely while keeping your local environment lightweight.
Customizing tools. To add your own tools to the MCP server, define each tool inside the init() method of your server’s source code using the pattern this.server.tool(...). This approach lets you expose new capabilities without changing client configurations.
Connecting via the Cloudflare AI Playground. You can test and experiment with your remote MCP server directly in the Cloudflare AI Playground by entering your deployed server URL. This lets you run tools through the playground’s remote MCP client and verify behavior before integrating into other clients.
Note on local development. The deployment flow described here focuses on a remote server without authentication. If you need authentication or a different hosting setup, you can adjust the deployment template or client configurations accordingly. Always ensure your server endpoint is accessible from your MCP clients and that you follow security best practices for exposing tooling endpoints.
Available tools
calculator
A calculator tool exposed via the MCP proxy. It is accessible remotely through the configured MCP server and can be invoked from compatible MCP clients.