Firecrawl MCP

API-based web scraping MCP server for AI agents.

What is this?

20 tools for web scraping, crawling, search, and data extraction — powered by Firecrawl's API. No proxy rotation. No anti-bot complexity. Just API calls.

Positioning

For AI agents that need web data without anti-bot headaches.

Unlike traditional scraping tools that require proxy rotation, browser stealth, and constant maintenance against anti-bot detection, Firecrawl MCP works through a simple API. The heavy lifting happens server-side.

Tools

Scraping

• scrape_and-extract-from-url — Scrape a single URL and return structured data
• batch_scrape-and-extract-from-urls — Batch scrape multiple URLs

Crawling

• crawl_urls — Crawl a website with configurable depth and scope
• crawl_get-status — Get crawl status
• crawl_cancel — Cancel a crawl job
• crawl_errors_get-crawl — Get crawl errors
• crawl_get-active — Get all active crawls

Search

• search — Search for URLs matching a query
• firecrawl-search_search-and-scrape — Search and scrape combined

Extraction

• extract_data — Extract structured data from URLs using selectors
• extract_get-status — Get extract job status

Deep Research

• deep-research_start — Start deep research job
• deep-research_get-status — Get deep research status

Maps

• map_urls — Generate a URL map for a website

LLM TXT

• llmstxt_generate-llms-txt — Generate llms.txt for a site
• llmstxt_get-llms-txt-status — Check llms.txt generation status

Team

• team_get-credit-usage — Get team credit usage
• team_get-token-usage — Get team token usage

Utilities

• context — Get API domain context
• sync — Sync operation
• export — Export data
• import — Import data
• sql — SQL query
• workflow_status — Get workflow status
• workflow_archive — Archive workflow

Installation

Apify

$apify push firecrawl-mcp

Local development

npm install
npm run build

Environment variables

Variable	Description
FIRECRAWL_BEARER_AUTH	Firecrawl API bearer token

Usage

Standalone (batch input)

{
  "tool": "scrape_and-extract-from-url",
  "params": {
    "url": "https://example.com"
  }
}

MCP Protocol (Standby mode)

Send JSON-RPC 2.0 requests to /mcp:

# Initialize
curl -X POST http://localhost:3000/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"test","version":"1.0"}}}'

# List tools
curl -X POST http://localhost:3000/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":2,"method":"tools/list","params":{}}'

# Call tool
curl -X POST http://localhost:3000/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"scrape_and-extract-from-url","arguments":{"url":"https://example.com"}}}'

Architecture

Apify Actor (handleRequest)
    |
    v
MCPProxy (Node.js child_process)
    |
    v
firecrawl-pp-mcp binary (stdio)
    |
    v
Firecrawl API

• Actor spawns firecrawl-pp-mcp as a subprocess with stdio transport
• JSON-RPC requests are proxied through stdin/stdout
• PPE charges applied via Actor.charge() before tool calls
• Standby HTTP server handles MCP protocol over HTTP

PPE Pricing

Tool	Price (USD)
scrape_and-extract-from-url	$0.10
batch_scrape-and-extract-from-urls	$0.10
crawl_urls	$0.15
map_urls	$0.05
search	$0.08
extract_data	$0.10
deep-research_start	$0.12
firecrawl-search_search-and-scrape	$0.08
context	$0.01
sync	$0.02
export	$0.03
import	$0.03
sql	$0.05
workflow_status	$0.02
workflow_archive	$0.03

Authentication

Firecrawl uses bearer token authentication. Set FIRECRAWL_BEARER_AUTH in Apify secrets.

Key Differentiators

1. No anti-bot — API-based scraping means no proxy rotation, no browser fingerprinting, no CAPTCHAs
2. No API key required — Bearer token auth via environment variable
3. 20 tools — Covering scrape, crawl, search, extract, and utilities
4. PPE ready — Per-tool pricing via Apify PAY_PER_EVENT model

GitHub Topics

firecrawl web-scraping ai-agents no-api-key-required mcp apify

Related

• Firecrawl Documentation
• Apify MCP Documentation
• Printing Press Library

License

MIT