firecrawl

What problem does it solve? Gathering content from the public web for research or data extraction is slow and error-prone when done manually, especially when pages require JavaScript rendering, pagination, or login. This Skill gives an AI agent a structured command-line workflow to search, scrape, map, crawl, and drive a cloud browser, returning clean markdown optimized for LLM context. ## Core Features & Use Cases - Escalating extraction workflow: Start with search to find pages, scrape a known URL, map to locate subpages, crawl for bulk site sections, and browser for interactive pages requiring clicks, form fills, or authenticated sessions. - File-based output isolation: Results are written to a .firecrawl/ directory with -o and read incrementally with grep or head, protecting the context window and treating fetched content as untrusted data. - Use Case: Research a competitor's documentation by searching for the docs domain, mapping it to find the authentication page, scraping that page to markdown, and using a named browser profile to log in and extract content behind authentication. ## Quick Start Ask the agent to scrape a specific public URL into markdown, for example: use firecrawl to scrape https://docs.firecrawl.dev and save the result to a local file.