Try “Veo”, “voiceover”, “thumbnail” or “Descript” · Esc to close

Firecrawl MCP Server 🔌 MCP server Freemium

Official Firecrawl server to scrape, crawl, map and search the web as clean Markdown

Research & web data MCP servers · Free plan, paid from $19/mo ★ 7.5k · MIT · updated 2026-09-25

8.6editor score
View on GitHub ↗
GitHub stars
7.5k
Stars this week
–
Forks
894
Licence
MIT
Last push
2026-09-25
Maintainer
Firecrawl
Installnpx -y firecrawl-mcp

Third-party MCP servers run with your permissions and API keys. Read the source before installing, and prefer pinned versions.

Works with

Built on the open MCP client standard, so it works in every app that supports MCP client.

Good for

About Firecrawl MCP Server

What it is

The Firecrawl MCP Server is Firecrawl's official open-source server for web data. It turns pages into LLM-ready Markdown or structured data and gives agents search, crawl and site-map tools. You can run it locally with npx or use Firecrawl's hosted endpoint, which has a keyless free tier.

Key features

  • Scrape single URLs to clean Markdown
  • Map a site's URLs and crawl multiple pages
  • Web search with page content
  • Autonomous research agent tool
  • Document parsing and page monitoring
  • Up to 26 tools at full configuration, including credit usage checks

How to install

Local: run npx -y firecrawl-mcp with FIRECRAWL_API_KEY set (or FIRECRAWL_API_URL for a self-hosted Firecrawl). Hosted: add https://mcp.firecrawl.dev/v2/mcp as a remote server. The README covers Cursor, VS Code, Windsurf and Claude Desktop.

Pricing and rights

The server is MIT-licensed. As of September 2026 Firecrawl's free plan gives 1,000 credits a month (about 1,000 scraped pages or 500 searches); Hobby costs $19 per month for 5,000 credits. You are responsible for respecting sites' terms and copyright when you reuse scraped text.

Who it is for

Writers, SEO teams and researchers who feed source pages, competitor content or documentation into an assistant. It also works for building source packs for AI-assisted articles, where the agent needs readable text rather than raw HTML.

Verdict

Mature, well-maintained and fast to set up, with a free hosted tier. Credits go quickly on crawls, and scraped content still needs checking for rights before you publish anything based on it.

web-scraping research markdown official open-source

Pros

  • Clean Markdown output
  • Hosted keyless free tier
  • Crawl, map, search and research in one server
  • Large, active project

Cons

  • Crawls burn credits fast
  • Scraped content brings copyright and terms questions
  • Many tools can crowd the context window

Exa MCP Server 🔌 MCP serverFreemium

Exa's hosted neural search server with web search, page fetch and deep research

★ 5.0k

Tavily MCP Server 🔌 MCP serverFreemium

Tavily's search, extract, map and crawl tools for agents, hosted or local

★ 2.4k

YouTube Transcript MCP 🔌 MCP serverOpen source

Community server that fetches YouTube captions, with optional TwelveLabs video analysis

★ 597

Popular searches