Firecrawl plugin

Firecrawl's own plugin for Claude. It connects Claude to Firecrawl's hosted search MCP server and adds skills that choose the right tool for each job: searching the web, reading the pages your searches find, looking up structured data from Firecrawl's Alexandria data providers, and researching developer docs, US government publications and scientific papers. What its skills cover - Web searches summarized with citations, and reading any page you name - Official library, SDK and API docs, plus code, issues and pull requests in public repositories - Scientific papers: find them, follow citations and read the passages that answer your question - US statutes, regulations, court opinions and other government publications - A company's existing logo, colours and fonts, or a website's design system written up as DESIGN.md - Published business contacts, lead lists, and enrichment for named professionals or a lead CSV - Current job openings, fantasy sports roster research and SEO audits of named pages - Deep research across many sources, ending in a cited answer or report Commands in Claude Code It adds /firecrawl:search, /firecrawl:scrape, /firecrawl:map, /firecrawl:crawl, /firecrawl:agent and /firecrawl:setup. Plain requests work too; the skills pick the tools. How to install it In Claude Code, open /plugin and install firecrawl from the official marketplace, or add Firecrawl's own marketplace with /plugin marketplace add firecrawl/firecrawl-claude-plugin, then /plugin install firecrawl@firecrawl. Choose one of the two. In Claude.ai or Cowork, upload the plugin ZIP under Customize > Plugins, then connect Firecrawl from the plugin's Connectors tab. How it connects It uses Firecrawl's hosted search MCP server at https://mcp.firecrawl.dev/v2/mcp-search. Run /mcp, pick firecrawl and sign in to Firecrawl in your browser. You never paste an API key into the chat. Good to know - Web, developer and paper searches and page reads use your Firecrawl credits; government search and finding data providers are free. - Requests send your search queries, page URLs and provider inputs to Firecrawl. - Version 2 no longer uses the Firecrawl CLI. The CLI's skills (crawl, interact, monitor and more) now come from Firecrawl's separate skills catalog, and Firecrawl MCP Server, listed separately, has the full scrape, crawl and browser tool set.

• “Find recent sources on remote work stipend policies and summarize them with citations.”

  1. firecrawl_search
  Copy

• “Read https://example.com/pricing and pull out each plan's price.”

  1. firecrawl_scrape
  Copy

• “Find papers on CRISPR off-target effects and the passages on how they're detected.”

  1. firecrawl_research_search_papers
  2. firecrawl_research_read_paper
  Copy

• “Which US federal regulations cover flying drones over people?”

  1. firecrawl_gov_search
  Copy

• “Document example.com's design system in DESIGN.md.”

  Copy

• “Find current remote openings for senior data engineers.”

  Copy