Model Context Protocol

Firecrawl MCP Server

Web Scraping & Search via Model Context Protocol

Give your AI agents the ability to browse the web and extract structured data. Firecrawl MCP provides high-quality markdown conversion for any URL.

Research and competitive intelligence teams use Firecrawl MCP to give their AI agents live internet access. Instead of manually Googling and copying text, your AI can now autonomously research competitors, gather market data, scrape documentation, and monitor news—all while converting cluttered HTML into clean Markdown with noise removed. This turns hours of manual research into 30-second automated intelligence.

  • ✓AI can browse live web content without manual copy-paste
  • ✓Structured markdown output, not messy HTML
  • ✓Handles JavaScript-rendered pages via headless browser
  • ✓Batch crawl entire domains for research
Deploy Managed ServerExplore Integrations
mcp-config.json
{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": [
  "-y",
  "firecrawl-mcp"
],
      "env": {
        "FIRECRAWL_API_KEY": "your_firecrawl_api_key"
      }
    }
  }
}
Real-World Automation

Common Workflows

See how teams combine this MCP with other tools to automate real business processes.

Competitor Intelligence Brief

Scenario: CEO asks: 'What did our top 3 competitors launch last month?' AI researches and compiles a summary.
Steps:
  1. AI identifies competitors: Acme, Globex, Initech
  2. Firecrawl MCP searches 'Acme new features March 2025' → scrapes top 5 results
  3. Repeat for each competitor, extracting announcements and blog posts
  4. Notion MCP writes summary to 'Competitive Intelligence' database
  5. Linear MCP creates task: 'Add competitor features to roadmap review'
Outcome: Competitive research that took 2 hours manually now completes in 90 seconds. Updated weekly automatically.

Lead Enrichment from Website

Scenario: New lead signs up with company 'acmecorp.com'. AI scrapes their website to understand their business before the sales call.
Steps:
  1. Firecrawl MCP crawls acmecorp.com (homepage, about, products, news)
  2. Extracts: industry, size, product descriptions, recent announcements
  3. AI summarizes: 'B2B SaaS, 50-100 employees, recent $20M Series B'
  4. HubSpot MCP updates lead record with enriched company profile
  5. Linear creates 'pre-call research done' task and assigns to rep
Outcome: Sales reps enter every call with deep context. Lead qualification accuracy improves, close rates up 18%.

Documentation QA

Scenario: After releasing v2.0, AI checks your docs to ensure all new features are documented correctly.
Steps:
  1. Firecrawl MCP crawls your docs site (docs.acme.com) collecting all pages
  2. AI scans for keywords like 'batch processing', 'multi-region' (new features)
  3. Flags missing or outdated pages
  4. Notion MCP creates task list: 'Update batch processing guide'
  5. Slack MCP posts to #docs: 'QA found 3 outdated pages'
Outcome: Documentation stays in sync with releases automatically. Reduced support tickets from missing docs by 40%.
Protocol Definition

Available Tools — In Depth

Detailed reference for each tool exposed by this MCP server, with examples and related use cases.

⚒

scrape_url

Convert any public URL (or authenticated page, if configured) into clean, structured Markdown. Firecrawl renders JavaScript, removes navigation/ads/footers, and extracts the core content with proper headings. Returns markdown with headings, lists, tables, and images as markdown links. Ideal for reading docs, blog posts, or news articles.

Example:""Scrape https://docs.firecrawl.dev/overview and summarize the key features in 3 bullet points.""
Works great with:search_webcrawl_site
⚒

search_web

Perform a Google-style web search and return the top 5-10 results with title, URL, and snippet. Uses Firecrawl's search API which aggregates multiple search engines. Use this to research current events, find documentation, or discover recent articles.

Example:""Search for 'best practices MCP server 2025'. Return top 5 results with URLs.""
Works great with:scrape_urlcrawl_site
⚒

crawl_site

Recursively crawl an entire domain (or subdomain) and extract all accessible pages. Returns a structured map of site content with URLs and extracted text snippets. Use this for comprehensive research, SEO analysis, or building a knowledge base from a docs site.

Example:""Crawl https://stripe.com/docs and return all pages related to 'refunds'.""
Works great with:scrape_urlsearch_web
Setup Guide

Configuration & Best Practices

Setup Checklist

  • ✓
    FIRECRAWL_API_KEY (required)
    Sign up at firecrawl.dev → Dashboard → API Keys. Copy the 'API Key' (fc-*).
  • ✓
    Credit balance
    Firecrawl uses credit-based pricing. Check balance in dashboard. Scrape costs ~1 credit/search hit; crawl costs per page.
  • ✓
    Rate limits
    Free tier: 5 credits/month. Paid: 10K-1M+ credits/month. Each tool call consumes credits. Monitor usage.
  • ✓
    Domain whitelisting (optional)
    For authenticated sites, configure allowed domains in Firecrawl dashboard → 'Authorized Domains'.
  • ✓
    Region selection
    Can specify region for search results (us for US-centric results). Default global.

Troubleshooting

401 Invalid API key
Fix: Key incorrect or deactivated. Regenerate in Firecrawl dashboard. Key format: fc-xxxxxxxxx
429 Insufficient credits
Fix: Plan exhausted. Top up credits or upgrade plan in dashboard.
403 Domain not whitelisted
Fix: Attempting to scrape an authorized domain. Add domain to 'Authorized Domains' in Firecrawl settings.
Timeout / site unreachable
Fix: Target site blocked Firecrawl's IP or requires auth. Try again later or use alternative source.
Rate limit: Depends on plan. Typically 5-20 requests/minute free; higher on paid. 429 on over-limit.

When to Use Firecrawl MCP Server vs. Alternatives

Use This MCP When:

  • You need AI-native access via natural language
  • Your workflows span multiple tools (MCP composability)
  • You prefer cloud hosting over local Docker
  • You want zero-config deployment with ClawFast
  • Your use case requires LLMs to reason and act autonomously

Consider Alternatives When:

  • You need bulk data sync (use native export/import)
  • Real-time streaming is critical (use native webhooks)
  • You have strict compliance requiring direct API audit logs
  • Your integration is a one-off script (direct SDK may be simpler)
  • You need features not yet exposed by this MCP server
FAQ

Common Questions About Firecrawl MCP Server

Q: What sites can I scrape?
A: Any publicly accessible website. For paywalled or login-required sites, use Firecrawl's authenticated爬虫 feature (enter session cookies) to scrape as logged-in user.
Q: JavaScript-heavy sites?
A: Firecrawl uses headless browser (Puppeteer) to render JS-heavy SPAs. Returns fully rendered DOM, not raw HTML.
Q: Anti-bot detection?
A: Firecrawl rotates IPs and fingerprints to avoid blocks. Some aggressive sites (e.g., LinkedIn) may still block—consider official APIs instead.
Q: Output format?
A: Structured Markdown with headings preserved, navigation/ads stripped, tables converted to markdown format, images as links.
Q: Credit cost per action?
A: search_web: ~1-3 credits per query. scrape_url: ~5-10 credits per page (more if JS-heavy). crawl_site: ~10-50 credits per 100 pages depending on complexity.
Q: Legal/ethical scraping?
A: Firecrawl respects robots.txt by default. Check site's TOS. Use for research, not mass redistribution. Rate-limited automatically.
Q: Can I scrape PDFs?
A: Firecrawl doesn't parse PDFs directly—use a PDF-to-text tool first or use other MCPs specialized in document processing.
Q: How current is the data?
A: Real-time: you scrape the live site. Cached results available for 24 hours if re-scraping same URL without changes.
Customer Success

Success Story: MarketPulse (market research firm)

Challenge

Analysts spent 6 hours/day manually visiting 50+ news sites, blogs, and social media to track product launches and pricing changes for client reports.

Solution

Deployed Firecrawl MCP + Notion MCP. Built 'Daily Intelligence' agent that crawls 100+ monitored sites at 6 AM, extracts relevant articles, writes summaries to Notion database, and posts highlights to Slack.

Results

  • ✓ Research time: 6 hrs/analyst/day → 15 min (98% reduction)
  • ✓ Coverage increased from 30 sites to 100+ without extra cost
  • ✓ Reports now include direct source links for verification
  • ✓ Team of 5 analysts now handles 3× client load
"We've scaled our intelligence operation without hiring. The AI does the stale research work; our analysts now focus on insight and strategy—the actual value-add work."
— Tomás Blanco, Senior Analyst, MarketPulse

Combine with Other MCPs

This MCP works great alongside other tools. Here are popular combinations.

slack
Post daily research summaries and breaking news alerts
→
notion
Store scraped content in Notion knowledge base for long-term reference
→
hubspot
Enrich lead profiles with company info from their website
→
linear
Create tasks to follow up on competitor product changes detected via scraping
→

Ready for full workflows?

Check out our integration page for Firecrawl MCP Server to see complete AI agent templates and step-by-step guides.

View Firecrawl MCP Server Integrations →

Technical Specifications

Protocol
SSE (Server-Sent Events)
Transport
HTTPS (TLS 1.3)
Authentication
Bearer token (Firecrawl API Key fc-*)
Rate Limit
Plan-dependent (5-100+ requests/min typical)
Typical Latency
1-5s per scrape (includes browser rendering)
Data Format
Markdown (clean) + JSON metadata
Supported LLMs
Claude 3.5+, GPT-4o, Kimi K2.5
Hosting
Firecrawl cloud (managed scraping infrastructure)
Uptime SLA
99.5%+ (Firecrawl status page)
Last Updated
On-demand per scrape (fresh every call)

Why Managed MCP?

Model Context Protocol (MCP) is the new standard for AI connectivity. While you can host servers locally, ClawFast provides a production-grade managed environment.

  • 24/7 Availability: No need to keep your local machine running.
  • Secure Secrets: API keys are encrypted at rest and never exposed to logs.
  • SSE Protocol: Native Server-Sent Events (SSE) support for cloud-to-cloud connectivity.
  • Zero Config: One-click deployment with pre-configured templates.

Looking for more ways to use Firecrawl MCP Server?

Explore our high-level integration page for Firecrawl MCP Server to see business use cases and ready-to-use AI agent templates.

View Firecrawl MCP Server Integrations →

Connect Firecrawl MCP Server to your AI stack today

Deploy your managed MCP server in under 60 seconds.

Start Free Deployment