Firecrawl

Firecrawl is a web scraping and crawling service that turns websites into LLM-ready data via scraping, crawling, mapping, extraction, and screenshotting. Exposes an MCP server so agents can call its tools without manual API setup. Used in the cluster as the default scraping primitive inside Claude Code agentic workflows, including lead-gen and job-listing extraction demos.

Key facts

• Free tier: 500 credits on sign-up, sufficient for small agentic demos
• Exposes an MCP server so Claude Code can call scrape/crawl/map/extract without wrapping API calls
• Returns structured markdown or JSON from arbitrary URLs
• Used in Herk’s dentist-leads and remote-jobs scraping demos
• In Roberts’s memory workflow, Firecrawl is used to gather deep web research for an Expert Knowledge Index, then feed that research into Pinecone or a local markdown memory system ^[src-059].

Related

• See also: Expert Knowledge Index, Pinecone, NotebookLM, Model Context Protocol (MCP)

Source references

• ^[src-005] Nate Herk cluster — Nate Herk — n8n cluster (18 videos)

– Videos referenced: 3GAxd90fEE4, AO5aW01DKHo

• ^[src-059] Jack Roberts — “This Memory System just 10x’d Claude Code” (2026-05-03)