Descriptions:
This tutorial from Nate Herk demonstrates how to connect Firecrawl’s MCP server to Claude Code inside VS Code, creating a reusable scraping agent that can extract structured data from any website in seconds. The video covers Firecrawl’s full range of endpoints — page scraping, site mapping, full crawling, and structured search — and explains why using the MCP interface lets Claude select the right tool automatically based on natural language instructions, rather than requiring manual API call construction.
Nate walks through the complete setup process: creating a new project in VS Code, securely storing the Firecrawl API key in a `.env` file to keep it out of conversation history, initializing the MCP connection through Claude Code, and generating a project-specific `CLAUDE.md` system prompt alongside a Firecrawl cheat sheet that Claude can reference when choosing between scrape, map, and crawl operations.
A practical end-to-end example shows the agent crawling a remote jobs site with roughly 1,700 listings spread across dozens of pages, extracting all results as structured data ready for export to a Google Sheet. The video also covers output formats including markdown, HTML, AI-generated summaries, full-page screenshots, and brand asset extraction. Viewers familiar with Claude Code basics will gain a reproducible workflow for turning any public website into LLM-ready data for downstream RAG pipelines or automation tasks.
📺 Source: Nate Herk · Published February 11, 2026
🏷️ Format: Tutorial Demo
