ScrAPI MCP Server is an MCP server that enables you to use ScrAPI to scrape web pages within your workflows. It exposes two tools: scrape_url_html and scrape_url_markdown. Both take a URL as input and return the scraped content in the requested format—HTML for structure-preserving results and Markdown for plain text extraction. These tools are designed for pages that are difficult to access due to bot detection, captchas, or geolocation restrictions. The MCP server can operate with or without an API key; without a key you’ll be limited to one concurrent call and twenty free calls per day. Cloud MCP endpoints are available at https://api.scrapi.tech/mcp/sse (SSE) and https://api.scrapi.tech/mcp (streamable HTTP), and you can test it with MCP Inspector. There is currently no facility to pass through your API key when connecting to the cloud MCP server. Claude Desktop users can configure the server via Docker or NPX entries, and a local Docker build is provided. The project is MIT licensed.