⚡ Fast Agent MCP Web Search & Clean Markdown Extractor

Fast Agent MCP Web Search & Extractor is an ultra-fast, zero-browser web search and clean LLM-optimized Markdown extractor designed specifically for AI Agents, Cursor, Claude Desktop, LangChain, and RAG pipelines.

---

🌟 Key Features

• 🏎️ Sub-Second Speed: Operates entirely without heavy headless browsers (Puppeteer/Playwright), saving 90%+ RAM and compute costs.
• 🧹 RAG-Ready Clean Markdown: Automatically strips scripts, styles, advertisements, cookie banners, navigation bars, and footers. Returns clean GitHub-Flavored Markdown with YAML metadata (tokens, title, author).
• 🔌 Native Model Context Protocol (MCP): Connects directly into Claude Desktop or Cursor as an autonomous tool.
• 💰 Zero Paid API Requirement: Built on public, lightweight search feeds. Zero OpenAI/Perplexity/SerpApi key fees required.
• 📦 Direct URL & Query Modes: Supports broad web search extraction or focused batch scraping across URL arrays.

---

🚀 Quick Start / Input Parameters

Example Input (JSON)

{
  "query": "Model Context Protocol architecture and best practices",
  "maxResults": 3,
  "extractFullContent": true,
  "mcpProtocolResponse": false
}

Parameter Details

Field	Type	Default	Description
query	string	""	Search keywords to retrieve live results for.
urls	array	[]	Direct list of web page URLs to scrape and convert into Markdown.
maxResults	integer	5	Maximum number of search results to fetch (1–20).
extractFullContent	boolean	true	When true, scrapes the full page content into clean Markdown.
includeRawHtml	boolean	false	When true, includes sanitized HTML alongside Markdown.
mcpProtocolResponse	boolean	false	When true, wraps results into MCP CallToolResult JSON.

---

💻 Claude Desktop & Cursor MCP Configuration

You can connect this Actor directly to Claude Desktop or Cursor via the Apify MCP connector:

{
  "mcpServers": {
    "apify-web-search-extractor": {
      "command": "npx",
      "args": [
        "-y",
        "@apify/mcp-server",
        "--actor",
        "username/fast-agent-mcp-web-search-extractor"
      ],
      "env": {
        "APIFY_TOKEN": "YOUR_APIFY_API_TOKEN"
      }
    }
  }
}

---

📊 Example Output

{
  "title": "Introducing Model Context Protocol",
  "url": "https://example.com/article",
  "snippet": "An open standard for AI assistants to securely access data...",
  "estimatedTokens": 412,
  "markdown": "---\ntitle: \"Introducing Model Context Protocol\"\nurl: \"https://example.com/article\"\ntokens_approx: 412\n---\n\n# Model Context Protocol\n\nModel Context Protocol (MCP) is an open standard...",
  "success": true
}

---

🏷️ Pricing & Monetization

• Pay Per Run: $0.0005 per search & extraction run.
• Monthly Rental: $19.00 / month with unlimited developer runs.
• Trial: 50 free runs included.

---

🛡️ License

Apache 2.0. Built for the Autonomous Zero-Cost Money Factory ecosystem.