The Firecrawl MCP Server is a tool that implements the Model Context Protocol (MCP). It enables advanced web scraping for LLM clients like Cursor and Claude.
You can start using it by running npx @mcp/firecrawl or installing it via npm. Configuration of an API key and environment variables is required.
Its key features include powerful web scraping with JavaScript rendering, automatic retries with exponential backoff, and efficient batch processing with rate limiting. It also offers comprehensive logging, credit usage monitoring, and supports both cloud and self-hosted instances.
Typical use cases are scraping website content for data analysis, automating data collection for research, and extracting structured information from web pages using LLM capabilities.
It can be used on any website that permits scraping. Rate limits are in place but configurable. Credit usage is tracked by the server's monitoring feature, which can alert you at set thresholds.
{
"mcpServers": {
"firecrawl-mcp": {
"command": "npx",
"args": [
"-y",
"firecrawl-mcp"
],
"env": {
"FIRECRAWL_API_KEY": "fc-af1b3ac1a0c2402485402fd0e34da158"
}
}
}
}