Appearance
Features Overview
Updated Feb 2026Firecrawl provides comprehensive web data extraction, browser automation, and AI-powered research capabilities.
Core Features
Feature Comparison
| Feature | Description | Cost | Best For |
|---|---|---|---|
| Scrape | Extract content from a single URL | 1 credit/page | Individual pages, targeted extraction |
| Batch Scrape | Scrape multiple URLs in parallel | 1 credit/page | Bulk extraction, efficiency |
| Search | Search web across sources/categories | 2 per 10 results | Finding information, research |
| Map | Discover all URLs on a site | 1 credit/call | Site structure, planning crawls |
| Crawl | Recursively crawl entire websites | 1 credit/page | Full site extraction, backups |
| Extract | Multi-URL structured extraction | Dynamic | Cross-page data aggregation |
| LLM Extract | JSON extraction with schemas | +4 credits/page | Structured data from pages |
| Browser | Remote browser sessions with interaction | 2 credits/min | Login flows, pagination, forms |
| Agent | AI-powered autonomous deep research | Dynamic | Complex queries, no URLs needed |
| Change Tracking | Monitor pages for content changes | Free (+5 for JSON) | Competitor monitoring, price tracking |
| Document Parsing | Extract from PDF, Word, Excel | +1/PDF page | Document processing |
| Enhanced Mode | Anti-bot bypass with stealth proxy | +4 credits/page | Protected websites |
| Fast Scraping | Cache-based speed optimization | Free | Repeated scrapes |
| MCP Server | Connect to AI coding tools | N/A | Claude Code, Cursor, VS Code |
All Features
Data Extraction
- Scrape → - Extract content in multiple formats
- Batch Scrape → - Parallel multi-URL scraping
- LLM Extract (JSON Mode) → - Schema-based structured extraction
- Extract API → - Multi-URL extraction endpoint
- Document Parsing → - PDF, Word, Excel processing
Web Operations
- Search → - Search the web with source/category filtering
- Map → - Discover website structure
- Crawl → - Recursive crawling with WebSocket watching
Browser & AI
- Browser Sandbox → - Remote browser sessions for interaction
- Agent → - AI-powered autonomous deep research
- FIRE-1 Model → - Browser-native agent for interactive tasks
- AI Models → - Spark-1 Mini, Pro, Fast + FIRE-1
Infrastructure
- Change Tracking → - Monitor content changes over time
- Enhanced Mode → - Anti-bot proxy and stealth mode
- Fast Scraping → - Cache-based speed optimization
- Proxies → - Standard, enhanced, and geographic targeting
- MCP Server → - Connect to Claude, Cursor, VS Code
See SOPs for step-by-step implementation guides or Developer Guides for integration patterns.