LinkedIn MCP Server

// mcp-server.js - CommonJS format for Netlify serverless function const { StreamableHTTPServerTransport } = require("@modelcontextprotocol/sdk").server.StreamableHTTPServerTransport; const { toFetchResponse, toReqRes } = require("fetch-to-node"); const { z } = require("zod"); const { McpServer } = require("@modelcontextprotocol/sdk").server.McpServer; const { JSONRPCError } = require("@modelcontextprotocol/sdk").types; // Netlify serverless function handler which handles all inbound requests exports.handler = async (event, context) => { const req = new Request(`https://${event.headers.host}${event.path}`, { method: event.httpMethod, headers: event.headers, body: event.body ? Buffer.from(event.body, 'base64').toString() : undefined }); try { // for stateless MCP, we'll only use the POST requests that are sent // with event information for the init phase and resource/tool requests if (req.method === "POST") { // Convert the Request object into a Node.js Request object const { req: nodeReq, res: nodeRes } = toReqRes(req); const server = getServer(); const transport = new StreamableHTTPServerTransport({ sessionIdGenerator: undefined, }); await server.connect(transport); let body; try { body = await req.json(); } catch (e) { console.error("Error parsing JSON body:", e); return { statusCode: 400, body: JSON.stringify({error: "Invalid JSON body"}) }; } await transport.handleRequest(nodeReq, nodeRes, body); nodeRes.on("close", () => { console.log("Request closed"); transport.close(); server.close(); }); const response = toFetchResponse(nodeRes); // Convert the Response to Netlify function format const responseBody = await response.text(); const headers = {}; response.headers.forEach((value, key) => { headers[key] = value; }); return { statusCode: response.status, headers: headers, body: responseBody }; } return { statusCode: 405, body: "Method not allowed" }; } catch (error) { console.error("MCP error:", error); return { statusCode: 500, headers: { "Content-Type": "application/json" }, body: JSON.stringify({ jsonrpc: "2.0", error: { code: -32603, message: "Internal server error", }, id: '', }) }; } }; function getServer() { // initialize our MCP Server instance that we will // attach all of our functionality and data. const server = new McpServer( { name: "basic-mcp-server", version: "1.0.0", }, { capabilities: { logging: {} } } ); // Add a simple tool for running an analysis report server.tool( "run-analysis-report", "Generates a sample data analysis report with random growth metrics.", { days: z.number().describe("Number of days to analyze").default(7), }, async ( { days }, ) => { const random = Math.random() * 100; return { content: [ { type: "text", text: JSON.stringify({ lastNDays: days, data: Array.from({ length: days }, (_, i) => `Day ${i + 1} had ${random * days} growth.`), }), }, ], }; } ); // Add a resource with information about interpreting reports server.resource( "interpreting-reports", "docs://interpreting-reports", { mimeType: "text/plain" }, async () => { return { contents: [ { uri: "docs://interpreting-reports", text: `Reports from this MCP include an array of text that informs the growth over a specified number of days. It's unstructured text but is consistent so parsing the information can be based on looking at a single line to understand where the data is.`, }, ], }; } ); return server; } // We don't need to export a config object here since we're using redirects in netlify.toml