firecrawl/firecrawl-mcp-server
🔥 Official Firecrawl MCP Server - Adds powerful web scraping and search to Cursor, Claude and any other LLM clients.
What it solves
Firecrawl MCP Server allows AI agents to interact with the live web. It solves the problem of agents being limited to their training data by providing tools to search, scrape, and interact with websites, turning raw web content into clean, structured, and agent-ready context.
How it works
It acts as a Model Context Protocol (MCP) server that connects AI agents to the Firecrawl API. It exposes a suite of tools that agents can call to perform web operations. These operations can be executed via a hosted cloud service or a self-hosted instance of Firecrawl.
Who it’s for
Developers building AI agents (using clients like Cursor, Windsurf, or Claude Desktop) who need their agents to have real-time web access, perform deep research, or extract structured data from websites.
Highlights
- Web Search & Discovery: Search the general web, specialized developer sources (GitHub issues, READMEs), or scientific research papers.
- Structured Extraction: Scrape URLs into clean JSON based on a provided schema or Markdown for full content.
- Website Mapping: Discover all indexed URLs on a specific site to plan scraping tasks.
- Browser Automation: Interact with pages by clicking, typing, and navigating.
- Autonomous Research: Use an integrated agent for complex, multi-source research tasks.
- Flexible Deployment: Supports hosted cloud endpoints, self-hosted instances, and multiple transport methods (stdio, SSE).
Related
- Project
- Project
- Project
- Project
- Project