firecrawl — MCP server for web scraping and crawling

draft — generated, polish on first use.

Category: pp-mcp Where it works: both Source: ~/.mcp.json

What it is

MCP server that exposes web scraping and crawling capabilities. Remote/Go implementation; tool surface not locally introspectable.

How to trigger it

  • “scrape this page”, “crawl the site”, “get content from URL”
  • “use firecrawl on
  • Auto-fires? no — explicit invocation required.

How to use it

Operator provides a URL or domain; firecrawl fetches and returns page content or crawl results. Output is raw or structured text depending on the chosen tool.

Practical examples

  1. you: “scrape https://example.com/pricing returns page text and metadata.
  2. you: “crawl docs site, depth 2” returns list of pages and extracted content.

Tools

toolwhat it doesexample
scrapefetches single page content and metadatascrape(url, formats=[“markdown”])
crawlrecursively crawls site and returns page listcrawl(url, maxDepth=2, limit=50)
mapreturns sitemap-like list of URLsmap(url)
extractstructured extraction via schema or promptextract(url, schema={…})
actionsbrowser actions (click, type, screenshot)actions(url, actions=[…])

Notes / limits

Remote service; verify actual tool availability at runtime. Rate limits and cost depend on upstream provider. If tool names differ from table above, verifică doctrina.