Firecrawl MCP Server

Give your AI agent the ability to search the web, scrape any webpage, crawl entire sites, and extract structured data. Works on JavaScript heavy single page apps that simple scrapers cannot handle.

MCP Server·4 sections·1 min read

researchmarketingcontent

Install Prompt

Paste this into Claude Code to set it up:

I want to add the Firecrawl MCP Server to my Claude Code setup.

IMPORTANT: Never use Read/Edit tools on ~/.claude.json. Use jq via Bash instead.

Prerequisites: You need a Firecrawl API key. Sign up at firecrawl.dev and grab your key from the dashboard. The free tier includes 500 credits.

Run this command to add the MCP server (replace your-api-key with your actual key):
```bash
jq '.mcpServers["firecrawl"] = {"type": "stdio", "command": "npx", "args": ["-y", "firecrawl-mcp"], "env": {"FIRECRAWL_API_KEY": "your-api-key"}}' ~/.claude.json > /tmp/claude-json-tmp && mv /tmp/claude-json-tmp ~/.claude.json
```

After adding, restart Claude Code. The server connects immediately and gives your agent the ability to search the web, scrape pages, crawl sites, and extract structured data.

01What It Does

The web, at your agent's fingertips

Firecrawl is to web content what GitHub MCP is to code: the connection that makes everything accessible. Unlike simple scrapers that break on modern JavaScript sites, Firecrawl renders pages fully before extracting content. It works on single page apps, dynamic dashboards, and anything that requires a real browser to load.

🌐

Search the WebRun search queries and get results with titles, URLs, and content snippets. Fast, cheap, and often enough on its own.

📄

Scrape Any PageFetch any webpage and get clean markdown output, stripping navigation, ads, and clutter. Handles JavaScript rendering.

🕸️

Crawl Entire SitesFollow links from a starting URL and extract content from every page. Great for docs audits and knowledge bases.

02Real Scenarios

"Research competitor pricing"

The agent searches for pricing pages, scrapes each one, and synthesizes findings into a comparison table with plans, prices, and features.

"Save this landing page to our swipe file"

You paste a URL. The agent scrapes it, analyzes the copy structure, extracts the value proposition, and saves organized notes.

"Find the latest pricing for Vercel"

The agent searches, finds the pricing page, scrapes it, and returns a clean summary of every plan with exact prices and limits.

"What are people saying about us online?"

The agent searches for your product name, scrapes relevant discussions, and summarizes sentiment, praise, and complaints.

03Pricing

	Price	Credits	Best For
Free	$0/month	500 credits	Trying it out, light personal use
Hobby	$16/month	3,000 credits	Regular research and content workflows
Standard	$83/month	100,000 credits	Teams and heavy scraping workloads
Growth	From $333/month	500,000+ credits	Production apps and large scale crawls

A single page scrape costs 1 credit. A search costs 1 credit per result page. Crawling costs 1 credit per page crawled. The most effective pattern is: search to find the right URLs, then scrape only the pages that matter.

04Works Well With

🔎

Perplexity MCPFirecrawl for raw content extraction, Perplexity for synthesized, sourced answers to complex questions. They complement each other perfectly.

📝

Notion MCPScrape web content with Firecrawl, then save structured research notes and summaries directly into Notion pages.

🎭

Puppeteer MCPFirecrawl reads page content. Puppeteer interacts with pages (clicking buttons, filling forms, taking screenshots). Different tools for different jobs.

✅

Todoist MCPResearch a topic with Firecrawl, then create actionable tasks in Todoist based on the findings.