Firecrawl MCP Server — The Full-Stack Web Scraping Platform for AI Agents

At a glance: 95,700+ parent repo stars, 5,800+ MCP server stars, 12+ tools, FIRE-1 web action agent (beta), Browser Sandbox, ~50.6K weekly PulseMCP visitors (#34 globally) Firecrawl isn't just a scraper — it's a web data platform with an MCP server. Where most web access tools give you one fetch endpoint, Firecrawl gives your agent an entire toolkit: scrape pages, crawl sites, search the web, extract structured data with LLM, run autonomous multi-source research, and control cloud browser sessions. What It Does Core scraping (5 tools): firecrawl_scrape — Single page to markdown, HTML, screenshots, or structured JSON firecrawl_batch_scrape — Parallel multi-URL scraping with rate limiting firecrawl_crawl — Async site crawling with depth control firecrawl_check_crawl_status — Monitor crawl progress firecrawl_map — Fast URL discovery without content extraction Search & extraction (2 tools): firecrawl_search — Web search with geo targeting and time filters firecrawl_extract — LLM-powered st