Firecrawl MCP Server

Example Use Cases

• 1

  Fetch and return the main content of a single web page as markdown or HTML (use scrape).

• 2

  Retrieve content from many known URLs efficiently with built-in rate limiting and parallel processing (use batch_scrape).

• 3

  Structured data extracted from pages via LLMs or want to discover site URLs before scraping (use extract and map/crawl).

Description

This MCP server integrates with Firecrawl to provide tools for single-page scraping, batch scraping, site mapping, full crawls, web search, and LLM-powered structured extraction. It supports cloud and self-hosted Firecrawl instances, configurable retry/backoff and credit monitoring, SSE and Streamable HTTP transport, and examples for integration with Cursor, Windsurf, VS Code, and Claude Desktop. Built-in rate limiting, automatic retries, and batch processing make large or parallel scraping jobs more reliable. The server returns content in markdown/html or structured JSON as appropriate and exposes job/status endpoints for long-running operations.