Introducing /interact. Scrape any page, then let your agent take over to click, type, and extract data for you. Try it now →
2 Months Free — Annually
Firecrawl in
Claude Code
Firecrawl is an official Claude plugin. Give Claude Code live web access — scrape, search, crawl, and extract from any website, right inside your coding workflow.
//
Used by over 500,000 developers
//
Trusted by 80,000+
companies of all sizes
companies of all sizes
[ 01 / 04 ]
·
Key Features
Your web context layer within Claude
Firecrawl turns any website into clean, LLM-ready data that Claude can actually use.
Scrape, crawl, search, map, extract - all within Claude Code
Handles JavaScript rendering, proxy rotation, and converts everything to clean markdown or structured JSON
Autonomous web extraction — just describe what you need and Firecrawl handles the rest
Perform web actions: click, scroll, write, wait, press and more before extracting content
Works with self-hosted Firecrawl instances as well
Scrape
Get LLM-ready data from any webpage. Markdown, JSON, screenshot, and more.
View docs
Search
Search the web and get full content from results in one call.
View docs
Agent
Gather data wherever it lives on the web with a single prompt.
View docs
[ 02 / 04 ]
·
Installation
Install Firecrawl in Claude Code
The best way to give Claude Code live web access. Pick the install method that works for you.
Option 1
Firecrawl Plugin
Install directly from the Claude plugin marketplace. Run /firecrawl:setup to add your API key.
1claude plugin install firecrawl@claude-plugins-officialOption 2
Firecrawl CLI + Skill
Install the CLI once and it auto-configures in Claude Code — better for token-heavy tasks.
1npx -y firecrawl-cli@latest init --all --browser
2
3# Your AI agent (Claude Code, Codex, OpenCode, etc) can now:
4# Scrape a page to clean markdown
5# Search and scrape top results
6# Crawl an entire website
7# Map an entire domain
8Option 3
Firecrawl MCP
Connect via the Model Context Protocol for inline results directly in your Claude Code session.
1claude mcp add firecrawl -e FIRECRAWL_API_KEY=your-api-key -- npx -y firecrawl-mcp[ 03 / 04 ]
·
Use Cases
Documentation Crawling
Pull live docs while you code
Don't let Claude hallucinate. Crawl any docs site and get current API references and guides directly into your session.
$ firecrawl crawl https://docs.stripe.com --limit 50 --wait
✓docs.stripe.com/get-started/setup
✓docs.stripe.com/payments
✓docs.stripe.com/payments/payment-intents
✓docs.stripe.com/api
✓docs.stripe.com/api/payment-intents
✓docs.stripe.com/webhooks
✓docs.stripe.com/testing
✓docs.stripe.com/billing
Autonomous Research Agent
Extract structured data on demand
Give Firecrawl a natural language prompt and it autonomously finds sources, navigates pages, and returns clean JSON.
Deep research in progress...
Academic papers
0 found
News articles
0 found
Expert opinions
0 found
Research reports
0 found
Industry data
0 found
Ask anything...
Competitive Research
Research competitors and monitor for changes
Scrape competitor pages, pricing tables, and feature lists — then track what changes over time without leaving your editor.
Competitor Monitor
Competitor A
competitor-a.com
3 changes
New pricing page
/pricing
2h ago
Feature update
/features
5h ago
Blog post
/blog/update
1d ago
Competitor B
competitor-b.com
1 change
Product launch
/products/new
3h ago
Content Generation
Generate content grounded in real web data
Scrape source pages into clean markdown so Claude's drafts are anchored to real content — not generic training data.
Content Pipeline
Source: techblog.com
→ Latest AI trends article
Generating...
5 articles created
SEO-optimized • Ready to publish
Social media posts
Twitter • LinkedIn • Instagram ready
[ 04 / 04 ]
·
What Our Customers Say
//
Community
//
People love
building with Firecrawl
Discover why developers choose Firecrawl every day.
Firecrawl is an open-source framework that takes a URL, crawls it, and conver..."
Upload a CSV of emails and..."
Firecrawl is an open-source framework that takes a URL, crawls it, and conver..."
Upload a CSV of emails and..."
Tutorials & guides
How to Create a Claude Code Skill: A Web Scraping Example with Firecrawl
Learn how to build a Claude Code skill that adds web scraping capabilities using Firecrawl. Covers skill structure, YAML frontmatter, and building multi-feature skills.
Read tutorial →
Top 10 Claude Code Plugins to Try in 2026
Discover the best Claude Code plugins that can transform your development workflow — from autonomous coding to browser automation and live web access with Firecrawl.
Read tutorial →
Claude Code for Marketing: How Marketers Can Use It in 2026
From shipping landing pages in hours to building custom agents for SEO workflows, here's how Claude Code and Firecrawl transformed a marketing team's day-to-day.
Read tutorial →
//
FAQ
//
Frequently
asked questions
Everything you need to know about Firecrawl in Claude Code.
General
Once installed, Firecrawl gives Claude Code six core tools: scrape (single pages), crawl (entire sites), search (web search with optional scraping), map (discover all URLs on a site), extract (AI-powered structured data extraction), and agent (autonomous research that finds sources and returns structured results from a single prompt). Claude will use these automatically when you ask about web content — or you can invoke them explicitly.
The Firecrawl Skill installs a command-line tool that Claude learns to use. Results are written to files on your filesystem rather than dumped directly into the context window — making it more token-efficient for large crawls. The Skill works across Claude Code, OpenAI Codex, and Gemini CLI with one install command. The MCP server connects Firecrawl directly via the Model Context Protocol and returns results inline inside Claude Code. Both give you the same Firecrawl capabilities — the Skill is recommended for most users.
Technical
Yes — a one-time restart is needed after running the install command so Claude Code can discover the new skill. After that, no restarts are required when using Firecrawl tools.
Yes. Set the FIRECRAWL_API_URL environment variable to point at your self-hosted instance (e.g., http://localhost:3002), and the skill will route all requests there instead of the cloud API. No API key is required for local instances.
Why Firecrawl?
The world's most comprehensive web data API. Our custom browser stack and semantic index deliver superior data quality across any website, handling more content types and edge cases than any competitor.
JavaScript rendering, dynamic content, and robust request handling built-in.
Process millions of pages with automatic rate limiting, caching, and distributed infrastructure.
Optimized scraping engine with parallel processing and smart caching for instant results.
Comprehensive docs, SDKs for all major languages, and dedicated support to help you succeed.
[ MAP ]
[ AGENT ]
[ SCRAPE ]
[ SEARCH ]
//
Get started
//
Ready to give Claude Code live web access?
Install the Firecrawl skill in 60 seconds and start building.
Are you an AI agent? Get an API key hereFOOTER
The easiest way to extract
data from the web
data from the web