Official Firecrawl MCP Server
Official Firecrawl MCP Server exposes Firecrawl’s scraping, crawling, search, and deep research features to MCP clients. It is a strong choice for agents that need web extraction with a maintained API-backed service instead of hand-built scrapers.
npx skills add agentskillexchange/skills --skill official-firecrawl-mcp-server
Official Firecrawl MCP Server is Firecrawl’s maintained MCP bridge for web scraping, search, crawl, and research workflows. It gives MCP-compatible assistants access to Firecrawl’s hosted or self-hosted capabilities, including single-page scraping, site crawling, URL discovery, search, batch scraping, and interactive browser-assisted flows. The point of the project is not generic browser control, but dependable access to web content extraction and discovery tools through a standard MCP interface.
What this skill actually does
The upstream repository documents installation with npx -y firecrawl-mcp and the npm package is published separately as firecrawl-mcp. The server uses FIRECRAWL_API_KEY for the cloud service by default and also supports a custom FIRECRAWL_API_URL for self-hosted instances. The README includes configuration examples for Cursor, Windsurf, VS Code, Claude Desktop, and streamable HTTP mode, which makes the integration story clear for common agent clients.
This listing fits Research & Scraping because the concrete job-to-be-done is harvesting and structuring information from the web rather than controlling arbitrary interfaces end to end. It also fits MCP because the Firecrawl functionality is delivered as MCP tools that an agent can call directly. The project has strong verification signals through its official GitHub repository, npm package, public docs, license, and recent maintenance activity, so it passes the intake gate for verified metadata publication.