Skip to main content
Piloterr
MCP logo

MCP & IDE agents

Call Piloterr from Cursor, Claude Code, and MCP clients

Two ways to connect: our hosted MCP at mcp.piloterr.com (Streamable HTTP) for Website Crawler and Rendering, or your own MCP server wrapping any REST endpoint from the Piloterr library.

  • Hosted MCP at https://mcp.piloterr.com/, crawler and rendering, no local process
  • Streamable HTTP, connect from Cursor, Claude Code, Gemini CLI, or ChatGPT Developer mode
  • Custom MCP server, expose LinkedIn, Amazon, Google News, or any library API
  • Same x-api-key header and credit pool for hosted and custom setups

At a glance

Hosted

mcp.piloterr.com

2

core scrape tools

HTTP

Streamable MCP

Custom

any library API

Why connect MCP

  • Hosted MCP | Streamable HTTP

    https://mcp.piloterr.com/ speaks MCP over Streamable HTTP. Add it as a remote HTTP server in your client with x-api-key, no local install. Website Crawler and Website Rendering are available as tools.

  • Not a web page

    Opening the URL in a browser tab will not work: only MCP clients send the required Accept headers to complete the handshake. Use Cursor, Claude Code, Gemini CLI, or ChatGPT Developer mode.

  • Custom MCP | full library

    Run a local MCP server (stdio) that calls any Piloterr REST endpoint, one @tool per API, or grouped tools your team reuses.

  • Production-grade bypass

    Both paths call Piloterr server-side, WAF, CAPTCHA, and JS rendering handled before JSON reaches the model.

Hosted vs custom MCP

What each path covers and when to use it.

  • Hosted: Website Crawler

    Fetch static HTML and JSON from URLs with advanced fingerprinting, ideal for quick page checks and link discovery from your editor.

  • Hosted: Website Rendering

    Headless Chrome renders JavaScript-heavy pages and returns clean content, use when crawler mode returns empty HTML.

  • Custom: Library endpoints

    Wrap /library APIs (LinkedIn, Amazon, Google News, etc.) as MCP tools, you choose which endpoints to expose.

  • Client compatibility

    Hosted: Cursor (Settings → MCP, HTTP streamable), Claude Code (claude mcp add --transport http), Gemini CLI (mcpServers httpUrl), ChatGPT Developer mode (remote MCP connector).

Why MCP + Piloterr beats raw fetch in agents

ApproachDIYPiloterr
Browser fetch toolBlocked on protected sitesManaged anti-bot bypass
Copy-paste cURLManual, not composableTyped MCP tools
HTML in contextToken-heavy, noisyCompact JSON fields
Ops overheadProxies, headless ChromeHosted or single API key

Connect MCP in four steps

  1. Step 1

    Create an API key

    Sign up for Piloterr, open the dashboard, and create an API key. Send it as the x-api-key header on every MCP request.

    Create an API key
  2. Step 2

    Choose hosted or custom

    Hosted: https://mcp.piloterr.com/, Website Crawler + Rendering only. Custom: a local server (stdio) for any library endpoint.

  3. Step 3

    Configure your MCP client

    Hosted, add an HTTP (streamable) server pointing to https://mcp.piloterr.com/ with x-api-key in headers (Cursor .mcp.json, Claude Code, Gemini CLI, or ChatGPT Developer mode). Custom, run FastMCP locally and reference the command in mcp.json.

  4. Step 4

    Verify and test

    Check server status at https://mcp.piloterr.com/health, then ask your agent to crawl a URL (hosted) or call a library tool (custom).

Workflow recipes

  • Hosted: crawl from Cursor

    Agent calls mcp.piloterr.com to fetch and render a competitor page while you write integration code.

  • Custom: library toolbox

    Local MCP server with your team's top library endpoints, LinkedIn, company enrich, SERP.

  • Claude Code + hosted MCP

    claude mcp add --transport http piloterr https://mcp.piloterr.com/ --header "x-api-key: …" for crawler tools in the terminal.

  • Hosted + custom together

    Remote crawler for ad-hoc URLs and a local server for structured library data in the same workspace.

Hosted vs custom MCP

  • Scenario

    Crawl or render any URL from the IDE

    Recommendation: Hosted MCP (Streamable HTTP)

  • Scenario

    LinkedIn, Amazon, or a specific library API

    Recommendation: Custom MCP server (stdio)

  • Scenario

    Scheduled production pipelines

    Recommendation: n8n or SDK

  • Scenario

    No-code SaaS triggers

    Recommendation: Zapier or Make

Hosted config and custom server

Remote Streamable HTTP server at mcp.piloterr.com, or a local stdio server for library APIs.

JSON
{
  "mcpServers": {
    "piloterr": {
      "url": "https://mcp.piloterr.com/",
      "headers": {
        "x-api-key": "your-x-api-key"
      }
    }
  }
}

Transparent credit pricing

Pay only for successful requests. Start with +500 credits, then scale with plans from $49/mo.

Premium

$49/mo

18,000 credits

Premium+

$99/mo

40,000 credits

Startup

$249/mo

110,000 credits

Ready to get started?

Your web scraping API is one click away. Start with +500 credits, no infrastructure to set up, no proxies to manage, and no credit card required.

Start free (+500 credits)