Firecrawl

HTTPv1.0.0Credential RequiredApr 11, 2026

Documentation

Guide

Firecrawl MCP Server

AI-powered web scraping, crawling, and extraction through the Model Context Protocol.

A Model Context Protocol (MCP) server that exposes Firecrawl's API for web scraping, website crawling, and structured data extraction.

Overview

The Firecrawl MCP Server provides stateless, multi-tenant access to:

Content Extraction: Scrape and extract markdown, HTML, JSON, and formatted content from single or multiple URLs
Website Crawling: Crawl entire websites with URL filtering, depth control, and scraping options
URL Discovery: Map all URLs on a website with filtering and geographic targeting
Web Search: Search the web and extract full content from results with advanced filtering
Structured Data Extraction: Extract data using JSON schemas and natural language prompts (async)
Autonomous Navigation: Use AI agents to navigate websites and complete multi-step tasks (async)

Perfect for:

Automated data collection and web intelligence pipelines
Building AI-powered research and analysis tools
Extracting structured data from unstructured web content
Real-time web monitoring and competitive analysis
Multi-step web automation and data enrichment tasks

Tools

Check server readiness and basic connectivity without authentication.

Inputs:

None

Output:

{
  "success": true,
  "status": "ok",
  "server": "CL Firecrawl MCP Server",
  "version": "0.1.0"
}

usage Example:

POST /mcp/health_check
 
{}

Scrape a single URL and extract content in multiple formats with flexible filtering options.

Inputs:

api_key (string, required) — Firecrawl API key for authentication
url (string, required) — The URL to scrape
formats (string, optional, default: "markdown") — Output formats (comma-separated): markdown, html, rawHtml, json, screenshot, links, images, summary, audio, branding, changeTracking
only_main_content (boolean, optional, default: true) — Extract only main content, excluding headers/footers/navs
include_tags (string, optional) — Comma-separated HTML tags to include in output
exclude_tags (string, optional) — Comma-separated HTML tags to exclude from output
wait_for_selector (string, optional) — CSS selector to wait for before scraping
(integer, optional, default: 30000) — Request timeout in milliseconds (1000-300000)

Crawl an entire website with URL filtering, discovery depth control, and comprehensive scraping options.

Inputs:

api_key (string, required) — Firecrawl API key for authentication
url (string, required) — The base URL to start crawling from
limit (integer, optional, default: 10000) — Maximum number of pages to crawl
sitemap (string, optional, default: "include") — Sitemap mode: skip, include, or only
max_discovery_depth (integer, optional) — Maximum depth based on discovery order
crawl_entire_domain (boolean, optional, default: false) — Follow sibling and parent URLs, not just child paths
allow_external_links (boolean, optional, default: false) — Allow crawler to follow external links
allow_subdomains (boolean, optional, default: false) — Allow crawler to follow subdomains

Get a complete list of all URLs from a website with optional filtering and geographic targeting.

Inputs:

api_key (string, required) — Firecrawl API key for authentication
url (string, required) — The base URL to start mapping from
limit (integer, optional, default: 5000) — Maximum number of URLs to return (max 100000)
search (string, optional) — Search query to filter and order results by relevance
sitemap (string, optional, default: "include") — Sitemap mode: skip, include, or only
include_subdomains (boolean, optional, default: true) — Include subdomains of the website
ignore_query_parameters (boolean, optional, default: true) — Don't return URLs with query parameters
ignore_cache (boolean, optional, default: false) — Bypass sitemap cache to get fresh URLs

Search the web and retrieve full page content from results with advanced geographical and source filtering.

Inputs:

api_key (string, required) — Firecrawl API key for authentication
query (string, required, max 500 characters) — Search query
limit (integer, optional, default: 5, 1-100) — Maximum number of results to return
sources (string, optional, default: "web") — Search sources (comma-separated): web, images, news
categories (string, optional) — Filter by categories (comma-separated): github, research, pdf
country (string, optional) — ISO country code for geo-targeting (e.g., US, DE, FR, JP)
location (string, optional) — Geographic location for search results (e.g., San Francisco,California,United States)
(string, optional) — Time-based search filter (e.g., qdr:d, qdr:w, qdr:m)

Start an asynchronous extraction job to extract structured data from multiple URLs using LLMs.

Inputs:

api_key (string, required) — Firecrawl API key for authentication
urls (string, required) — Comma-separated URLs to extract from (glob format supported)
prompt (string, optional) — Custom prompt to guide the extraction process
schema (string, optional) — JSON schema to structure the extracted data
enable_web_search (boolean, optional, default: false) — Use web search to find additional data during extraction
ignore_sitemap (boolean, optional, default: false) — Ignore sitemap.xml files during website scanning
include_subdomains (boolean, optional, default: true) — Include subdomains when scanning URLs
show_sources (boolean, optional, default: false) — Include sources used for extraction in response

Check the status of an extraction job and retrieve results when complete.

Inputs:

api_key (string, required) — Firecrawl API key for authentication
job_id (string, required) — The extraction job ID returned by the extract tool (UUID format)

Output:

{
  "success": true,
  "status": "completed",
  "data": {
    "extracted_data": [...]
  },
  "tokensUsed": 1250

Start an asynchronous autonomous AI agent to navigate websites and complete complex multi-step tasks.

Inputs:

api_key (string, required) — Firecrawl API key for authentication
prompt (string, required, max 10000 characters) — Natural language description of data to extract or task to complete
urls (string, optional) — Comma-separated URLs to constrain the agent to
schema (string, optional) — JSON schema for structured output
model (string, optional, default: "spark-1-mini") — Model to use: spark-1-mini (cheaper) or spark-1-pro (higher accuracy)
strict_constrain_to_urls (boolean, optional, default: false) — If true, agent will only visit URLs provided in urls parameter
max_credits (float, optional, default: 2500) — Maximum credits to spend on this agent task

Output:

Check the status of an agent job and retrieve results when complete.

Inputs:

api_key (string, required) — Firecrawl API key for authentication
job_id (string, required) — The agent job ID returned by the agent tool (UUID format)

Output:

{
  "success": true,
  "status": "completed",
  "data": {
    "extracted_data": {...}
  },
  "model": "spark-1-mini"

API Parameters Reference

api_key — Firecrawl API key for authentication (required for all tools except health_check)
timeout_ms or timeout — Request timeout in milliseconds (1000-300000ms)
formats — Output format types (comma-separated): markdown, html, rawHtml, json, screenshot, links, images, summary, audio, branding, changeTracking
mobile — Emulate mobile device (responsive layout)
proxy — Proxy type: basic, enhanced, or auto
block_ads — Block ads and cookie popups

Job IDs:

uuid format (e.g., "550e8400-e29b-41d4-a716-446655440000")

url Patterns:

single: https://example.com
Multiple (comma-separated): https://example.com/page1,https://example.com/page2
Glob patterns: https://example.com/products/*.html

status Values:

sync endpoints: success boolean (true

Authentication Guide

All tools require authentication via the Firecrawl API key.

Step 1: Get Your API Key

Go to Firecrawl Dashboard
Sign up or log in to your account
Navigate to API Keys section
Copy your active API key (starts with fc-)

Step 2: Authenticate in Tool Calls

Every tool (except health_check) requires the api_key parameter passed per request:

{
  "tool": "scrape",
  "arguments": {
    "api_key": "fc-YOUR-API-KEY",

Troubleshooting

Resources

Connect

Get connection URL

https://gateway.mewcp.com/firecrawl/mcp

Code Snippets

{
  "inputs": [
    {
      "type": "promptString",
      "id": "firecrawl-api-key",
      "password": true
    }
  ],
  "servers": {
    "firecrawl": {
      "type": "http",
      "url": "https://gateway.mewcp.com/firecrawl/mcp",
      "headers": {
        "Authorization": "Bearer ${input:firecrawl-api-key}"
      }
    }
  }
}

Request Builder

import asyncio
from fastmcp import Client

# Real MewCP gateway URL for this server
GATEWAY_URL = "https://gateway.mewcp.com/firecrawl/mcp"
API_KEY = "YOUR_API_KEY"

headers={
    "Authorization": f"Bearer {API_KEY}"
}

async def main():
    
    client = Client(GATEWAY_URL, headers=headers)
    
    async with client:
        # Test the connection by pinging the server
        await client.ping()
        
        # List available operations
        tools = await client.list_tools()
        resources = await client.list_resources()
        prompts = await client.list_prompts()

        # Execute operations
        result = await client.call_tool("example_tool", {"param": "value"})
        print(result)

asyncio.run(main())

Firecrawl

Documentation

Firecrawl MCP Server

Overview

Tools

API Parameters Reference

Authentication Guide

Step 1: Get Your API Key

Step 2: Authenticate in Tool Calls

Troubleshooting

Resources

Firecrawl MCP Server

Overview

Tools

API Parameters Reference

Authentication Guide

Step 1: Get Your API Key

Step 2: Authenticate in Tool Calls

Troubleshooting

Resources

Step 3: Stateless Multi-Tenancy

Step 4: Required Scopes