SSE

SSE servers can be run as is, we just need to identify the port used and adjust the command, and disable the smoke checks as they will confuse the server.

Start by cloning the MCP servers git repository and making a copy of the Dockerfile:

git clone https://github.com/modelcontextprotocol/servers.git
cd servers
cp src/everything/Dockerfile .

Make the following changes to the Dockerfile:

 RUN npm ci --ignore-scripts --omit-dev
+
+EXPOSE 3001

-CMD ["node", "dist/index.js"]
+CMD ["node", "dist/sse.js"]

Now run:

fly launch --ha=false --smoke-checks=false

Access the MCP server via the MCP inspector:

npx @modelcontextprotocol/inspector

In the top left, change the transport type to SSE.

Set the URL to match your application, where the URL will be of the form:

https://appname.fly.dev/sse

Replace the appname with the name of your application, but keep the https:// and /sse.

Click Connect.

An example Cursor configuration:

{
  "mcpServers": {
    "everything": {
      "url": "https://appname.fly.dev/sse",
      "env": {}
    }
  }
}

With SSE transports, fly mcp proxy may not be required, but could be useful if:

• Your MCP client doesn’t support SSE
• You need to set up a wireguard tunnel to access your MCP serever via an .internal or .proxy address.
• You need to route the request to a specific machine.