freshcontext-mcp

Instructions

Implementation Reference

• src/server.ts:98-118 (handler)

  The MCP tool registration and handler implementation for "extract_yc" in src/server.ts.

  server.registerTool(
    "extract_yc",
    {
      description:
        "Scrape YC company listings. Use https://www.ycombinator.com/companies?query=KEYWORD to find startups in a space. Returns name, batch, tags, description per company with freshness timestamp.",
      inputSchema: z.object({
        url: z.string().url().describe("YC companies URL e.g. https://www.ycombinator.com/companies?query=mcp"),
        max_length: z.number().optional().default(6000),
      }),
      annotations: { readOnlyHint: true, openWorldHint: true },
    },
    async ({ url, max_length }) => {
      try {
        const result = await ycAdapter({ url, maxLength: max_length });
        const ctx = stampFreshness(result, { url, maxLength: max_length }, "ycombinator");
        return { content: [{ type: "text", text: formatForLLM(ctx) }] };
      } catch (err) {
        return { content: [{ type: "text", text: formatSecurityError(err) }] };
      }
    }
  );

• src/adapters/yc.ts:5-103 (handler)

  The core logic for scraping YC data, implemented in ycAdapter.

  export async function ycAdapter(options: ExtractOptions): Promise<AdapterResult> {
    const safeUrl = validateUrl(options.url, "yc");
    options = { ...options, url: safeUrl };
  
    const browser = await chromium.launch({ headless: true });
    const page = await browser.newPage();
  
    // YC company directory is React-rendered — wait for network to settle
    await page.goto(options.url, { waitUntil: "networkidle", timeout: 30000 });
  
    // Wait for company cards to appear
    await page.waitForSelector('a[href*="/companies/"]', { timeout: 15000 }).catch(() => null);
  
    const data = await page.evaluate(`(function() {
      // YC company cards — robust multi-strategy extraction
      var results = [];
  
      // Strategy 1: structured company divs with name + description + batch
      var cards = Array.from(document.querySelectorAll('div[class*="_company_"]'));
  
      if (cards.length === 0) {
        // Strategy 2: anchor links to /companies/* pages
        cards = Array.from(document.querySelectorAll('a[href*="/companies/"]'))
          .filter(function(el) {
            return el.querySelector('span, p, div');
          });
      }
  
      cards.slice(0, 25).forEach(function(el) {
        var allText = el.innerText || el.textContent || "";
        var lines = allText.split('\\n').map(function(l) { return l.trim(); }).filter(Boolean);
  
        // Try to find structured spans
        var spans = Array.from(el.querySelectorAll('span'));
        var name = null, description = null, batch = null;
        var tags = [];
  
        spans.forEach(function(s) {
          var t = s.textContent.trim();
          if (!t) return;
          if (s.className && s.className.toString().includes('Name')) name = t;
          else if (s.className && s.className.toString().includes('Desc')) description = t;
          else if (s.className && s.className.toString().includes('Batch')) batch = t;
          else if (s.className && s.className.toString().includes('Tag')) tags.push(t);
        });
  
        // Fallback to line parsing
        if (!name && lines.length > 0) name = lines[0];
        if (!description && lines.length > 1) description = lines[1];
  
        var link = el.tagName === 'A'
          ? el.getAttribute('href')
          : (el.querySelector('a') ? el.querySelector('a').getAttribute('href') : null);
  
        if (name && name.length > 1 && name.length < 80) {
          results.push({ name, description, batch, tags, link });
        }
      });
  
      return results;
    })()`);
  
    await browser.close();
  
    const typedData = data as Array<{
      name: string | null;
      description: string | null;
      batch: string | null;
      tags: string[];
      link: string | null;
    }>;
  
    if (!typedData.length) {
      return {
        raw: "No YC companies found — page may have changed structure. Try visiting: " + options.url,
        content_date: null,
        freshness_confidence: "low",
      };
    }
  
    const raw = typedData
      .map((r, i) =>
        [
          `[${i + 1}] ${r.name ?? "Unknown"}`,
          `Batch: ${r.batch ?? "Unknown"}`,
          `Tags: ${r.tags?.join(", ") || "none"}`,
          `Description: ${r.description ?? "N/A"}`,
          `Link: ${r.link ? (r.link.startsWith("http") ? r.link : "https://www.ycombinator.com" + r.link) : "N/A"}`,
        ].join("\n")
      )
      .join("\n\n")
      .slice(0, options.maxLength ?? 6000);
  
    return {
      raw,
      content_date: new Date().toISOString().split("T")[0],
      freshness_confidence: "high",
    };
  }