MCP Web Tools Server

import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js"; import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js"; import { z } from "zod"; import { fetch } from "undici"; import { JSDOM } from "jsdom"; import Readability from "@mozilla/readability"; const server = new McpServer({ name: "mcp-web-tools-server", version: "0.1.0", }); const SEARCH_RESULTS_LIMIT_DEFAULT = 10; const MAX_FETCH_BYTES = 1024 * 1024 * 16; async function ddgSearch(query, limit) { const params = new URLSearchParams({ q: query, kl: "us-en" }); const res = await fetch(`https://duckduckgo.com/html/?${params.toString()}`, { headers: { "User-Agent": "Mozilla/5.0 (compatible; MCP-Web-Tools/0.1; +https://example.com)", "Accept-Language": "en-US,en;q=0.9", }, }); const html = await res.text(); const dom = new JSDOM(html); const doc = dom.window.document; const results = []; const nodes = doc.querySelectorAll(".result__title a.result__a"); for (const a of nodes) { const title = a.textContent?.trim() || ""; let href = a.getAttribute("href") || ""; if (!href) continue; let url = href; try { if (href.startsWith("/l/?") || href.includes("duckduckgo.com/l/?")) { const full = href.startsWith("http") ? href : `https://duckduckgo.com${href}`; const u = new URL(full); const target = u.searchParams.get("uddg"); if (target) url = decodeURIComponent(target); } else if (href.startsWith("//")) { url = `https:${href}`; } else if (href.startsWith("/")) { url = `https://duckduckgo.com${href}`; } } catch { } if (!url) continue; results.push({ title, url }); if (results.length >= limit) break; } return results; } async function serpApiSearch(query, limit) { const apiKey = process.env.SERPAPI_KEY; if (!apiKey) return null; const params = new URLSearchParams({ engine: "google", q: query, num: String(Math.min(limit, 10)), api_key: apiKey, }); const res = await fetch(`https://serpapi.com/search.json?${params.toString()}`); if (!res.ok) throw new Error(`SerpAPI error: ${res.status}`); const data = await res.json(); const organic = Array.isArray(data.organic_results) ? data.organic_results : []; return organic.slice(0, limit).map((r) => ({ title: r.title || "", url: r.link || "" })); } server.tool( "search_web", { query: z.string().min(1), limit: z.number().int().min(1).max(10).optional(), }, async (input) => { const limit = input.limit ?? SEARCH_RESULTS_LIMIT_DEFAULT; let results = null; try { results = await serpApiSearch(input.query, limit); } catch (e) { } if (!results) { results = await ddgSearch(input.query, limit); } const text = results .map((r, i) => `${i + 1}. ${r.title}\n${r.url}`) .join("\n\n"); return { content: [{ type: "text", text: text || "No results" }], }; } ); server.tool( "fetch_url", { url: z.string().url(), maxBytes: z.number().int().min(1024).max(16 * 1024 * 1024).optional(), }, async (input) => { const maxBytes = Math.min(input.maxBytes ?? MAX_FETCH_BYTES, 16 * 1024 * 1024); const res = await fetch(input.url, { headers: { "User-Agent": "Mozilla/5.0 (compatible; MCP-Web-Tools/0.1; +https://example.com)", Accept: "*/*", }, }); const contentType = res.headers.get("content-type") || ""; const reader = res.body.getReader(); let received = 0; const chunks = []; while (true) { const { done, value } = await reader.read(); if (done) break; received += value.byteLength; if (received > maxBytes) { chunks.push(value.subarray(0, value.byteLength - (received - maxBytes))); break; } chunks.push(value); } const body = Buffer.concat(chunks.map((u) => Buffer.from(u))).toString("utf8"); const summary = `contentType: ${contentType}\nbytes: ${Math.min(received, maxBytes)}`; return { content: [ { type: "text", text: summary }, { type: "text", text: body }, ], }; } ); server.tool( "extract_readable", { url: z.string().url() }, async (input) => { const res = await fetch(input.url, { headers: { "User-Agent": "Mozilla/5.0 (compatible; MCP-Web-Tools/0.1; +https://example.com)", }, }); const html = await res.text(); const dom = new JSDOM(html, { url: input.url }); const reader = new Readability(dom.window.document); const article = reader.parse(); if (!article) { return { content: [{ type: "text", text: "No readable content found." }] }; } const textBlocks = []; if (article.title) textBlocks.push(`# ${article.title}`); if (article.byline) textBlocks.push(`by ${article.byline}`); if (article.excerpt) textBlocks.push(article.excerpt); if (article.textContent) textBlocks.push(article.textContent); return { content: [{ type: "text", text: textBlocks.join("\n\n") }] }; } ); const transport = new StdioServerTransport(); await server.connect(transport);