Skip to content

MCP setupData and AI

Firecrawl MCP Server

Scrape pages, crawl whole sites and pull structured data into your agent as clean markdown with the official Firecrawl MCP server.

npx workspacetemplates add firecrawl-mcp --tool claude-code

Writes 1 file to your project root

.mcp.jsonmerge into your file14 lines · 213 BFile
{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": [
        "-y",
        "firecrawl-mcp"
      ],
      "env": {
        "FIRECRAWL_API_KEY": "${FIRECRAWL_API_KEY}"
      }
    }
  }
}

What it does

Firecrawl turns web pages into text an agent can use. Its MCP server can scrape a single URL, map a site's links, search the web, crawl many pages and extract structured fields, returning markdown instead of raw HTML.

This setup runs the npm package locally with your Firecrawl API key. Requests go to Firecrawl's service, so the pages you ask about are fetched by that service and each call uses credits on your plan. Firecrawl also offers hosted endpoints and a self-hosted option, described in its README.

Example

you › Scrape https://example.com/pricing and list each plan with its monthly price.

agent › 3 plans: Free at $0, Team at $24 per seat, Business at $60 per seat. Source: the pricing table, scraped as markdown.

Questions

Does it cost money?
Calls use credits from your Firecrawl account. Check your plan's allowance before asking for a large crawl.
Can it read pages behind a login?
Not by default. It fetches public pages. Never paste a session cookie into a prompt for it.
Is there a self-hosted option?
Yes. The README describes FIRECRAWL_API_URL for pointing the server at your own Firecrawl instance.
  • Apify MCP Server

    MCP setupData and AI

    Free
  • Chroma MCP Server

    MCP setupData and AI

    Free
  • ClickHouse MCP Server

    MCP setupData and AI

    Free
  • Hugging Face MCP Server

    MCP setupData and AI

    Free