ScrAPI MCP Server

What is ScrAPI MCP Server?

ScrAPI MCP Server lets MCP-compatible clients scrape web pages through ScrAPI, a service that handles real browser sessions, CAPTCHA solving, residential proxies, cookie banners, JavaScript rendering, geolocation-aware fetching, and pre-scrape browser actions. It is for developers who need to extract data from pages that require advanced browser automation.

How to use ScrAPI MCP Server?

Install via npx, Docker, or Smithery, and set the SCRAPI_API_KEY environment variable. Use the provided tools scrape_url_html and scrape_url_markdown with a target URL and optional browser commands. Configure your MCP client (e.g., Claude Desktop, Cursor, VS Code) with stdio or Streamable HTTP transport.

Key features of ScrAPI MCP Server

• Scrape any valid https:// or http:// URL through ScrAPI.
• Return results as raw HTML or readable Markdown.
• Execute browser commands (clicking, scrolling) before scraping.
• Supports stdio transport for desktop clients.
• Supports Streamable HTTP transport for remote clients.
• Run with npx, Docker, Smithery, or from source.

Use cases of ScrAPI MCP Server

• Extract article text or product copy for LLM analysis.
• Retrieve page structure with links, tables, and metadata.
• Scrape content from sites requiring CAPTCHA or residential proxies.
• Automate data collection from JavaScript-rendered pages.
• Enable AI coding assistants to fetch and analyze web content.

FAQ from ScrAPI MCP Server

What tools does the server provide?

It provides scrape_url_html (returns HTML) and scrape_url_markdown (returns Markdown). Both accept a URL and optional browser commands.

Do I need an API key?

Yes, SCRAPI_API_KEY is required. Without one, ScrAPI allows limited free usage on certain domains with lower concurrency and queueing priority.

How do I set the transport?

By default it uses stdio. Set TRANSPORT=http to use Streamable HTTP over a configurable port (default 5000).

What are the runtime requirements?

Node.js 18+ for npx or local development, Docker for containers, and an MCP-compatible client (e.g., Claude Desktop).

Where does the data go?

The server sends requests to ScrAPI's upstream API. Responses are returned directly to the MCP client. No local data storage is mentioned.