GTM Hub
What is Firecrawl?
Open-source web data API for AI agents and LLM applications. Managed API (no local browser needed) that searches, scrapes, crawls, parses, and interacts with the web at scale. Also available open-source for self-hosting.
Key features
Search
Live web queries returning full page content, not just links or snippets
Scrape
Turn any URL into clean Markdown or structured JSON. Handles JS-heavy pages, dynamic sites, and outputs AI-ready markdown with boilerplate, ads, and navigation stripped
Crawl
Scrape every subpage of a site automatically
Map
Discover all URLs on a website without scraping content
Parse
Document extraction from PDFs, DOCX, and other file formats
Interact
Managed browser automation for dynamic pages and multi-step flows (click, type, fill forms). Not designed for long-running persistent sessions
Agent
Describe the data you need via prompt and Agent searches, navigates, and returns results—URLs optional
Monitor
Recurring checks on named pages with webhook, email, or Slack notification when something changes
Firecrawl Alexandria
400+ data providers and specialized indexes (companies, people, finance, legal, real estate)
JavaScript rendering
All scrapes run in real Chromium, handling React, Vue, Next.js sites
Multiple output formats
Markdown, JSON, HTML, screenshots, links, images, branding, product extraction, summaries
Firecrawl pricing
Credit-based system. 1 credit = 1 page for basic scrape or crawl. Search costs 2 credits per 10 results. Structured formats (JSON, question, highlights) add 4 credits per page. PDF parsing adds 1 credit per page. Agent has 5 free runs daily, then dynamic pricing. Scraped pages cached on paid plans, free plans don't cache. Pay-as-you-go available on paid plans.
| Plan | Price | What's included |
|---|---|---|
| Free | $0/month | 1,000 credits/month (500 searches or 1,000 page scrapes), 2 concurrent browsers, low rate limits, no credit card required |
| Hobby | $19/month or $16/month annually | 5,000 credits/month (2,500 searches or 5,000 pages), 5 concurrent browsers, basic support |
| Standard | $99/month or $83/month annually | 100,000 credits/month (50,000 searches or 100,000 pages), 25 concurrent browsers, standard support, 500/min rate limit for scrape/search |
| Growth | $399/month or $333/month annually | 500,000 credits/month (250,000 searches or 500,000 pages), 50 concurrent browsers, priority support, 5,000/min rate limit |
| Scale | $749/month or $599/month annually | 1,000,000 credits/month (500,000 searches or 1,000,000 pages), 100 concurrent browsers, priority support, 10,000/min rate limit |
| Enterprise | Custom | Custom credits, concurrency, rate limits. Dedicated support, SLA, bulk discounts, zero-data retention, SSO, advanced security, multi-team organizations |
Pros & cons
Pros
- + Handles JavaScript-rendered content, login walls, and region-locked sites
- + AI-ready output by default (strips ads, nav, boilerplate; outputs clean Markdown)
- + Multiple integrations: works as Claude connector, ChatGPT plugin, MCP server for coding agents, REST API, SDKs
- + Flexible pricing: free tier with no card, pay-as-you-go on paid plans, credits roll over
- + P95 latency of ~3.4 seconds on 1,000-URL benchmark
- + Open-source and can be self-hosted
- + Structured data extraction without schema requirement
- + Built for AI agents: Agent endpoint, supports batch processing, integrates with MCP standard
Cons
- − Interact endpoint not suitable for long-running persistent browser sessions
- − Additional costs for structured formats (JSON extraction adds 4 credits per page)
- − Keyless tier has rate limits and no Alexandria access
- − Concurrent browser limits scale with plan; free tier limited to 2
- − Screenshots expire after 24 hours
Who is Firecrawl best for?
- • AI agents and LLM applications needing web data
- • Real-time data extraction from dynamic JavaScript-heavy sites
- • Batch scraping thousands of URLs in parallel
- • Structured data extraction (products, pricing, contact info)
- • Multi-step web automation and interaction workflows
- • Research and discovery tasks where URLs are not known upfront
- • Companies and teams already using Claude, ChatGPT, or MCP-based tools
