Firecrawl MCP

Firecrawl runs a hosted MCP server at https://mcp.firecrawl.dev/v2/mcp. Without a key it exposes firecrawl_search, firecrawl_scrape and firecrawl_parse, rate-limited. With OAuth or a Firecrawl API key the full set opens: firecrawl_map and firecrawl_crawl for whole sites, firecrawl_interact for pages that need a click or a form, firecrawl_agent for multi-source research, and firecrawl_monitor_* to watch a page for changes. Scrapes return clean markdown or JSON matching a schema you supply.

What it adds to Moby. Reading, not just finding. Moby surfaces the token and the wallets; Firecrawl turns the project's website, docs, GitHub and the articles about it into text your agent can quote, and can watch a page (a token's docs, a launchpad's announcements) for changes.

How to pair it. After get_token_overview and get_token_x_profile, scrape the project's site and whitepaper with firecrawl_scrape, and set a firecrawl_monitor on the page a thesis depends on. Run the result past get_token_risk_signals before anything is bought.

What it does not do. It reads public web pages; it has no on-chain, holder or wallet data and cannot trade. Crawls spend credits quickly, so bound them with limit. A page can say anything; verify contract addresses with Moby's search_tokens.