Firecrawl
An API that turns any website into clean, LLM-ready Markdown (or structured JSON). It handles the messy parts of web data — JavaScript rendering, pagination, anti-bot friction, boilerplate stripping — and gives back content an agent can actually use. Modes cover single-page scrape, whole-site crawl, map (discover URLs), search, and schema-driven extract.
Links
Description
- Scrape — one URL → clean Markdown/HTML/JSON, JS rendered.
- Crawl — follow links across a whole site, returning every page as Markdown.
- Map — quickly enumerate a site’s URLs.
- Search — web search with content fetched and returned.
- Extract — pull structured data against a schema (LLM-assisted).
- SDKs & integrations — Python/Node SDKs; works as an agent tool / MCP server.
Download or use
pip install firecrawl-py # or: npm i @mendable/firecrawl-js
# app.scrape_url("https://example.com", params={"formats":["markdown"]})- Site: firecrawl.dev
- Repo: github.com/mendableai/firecrawl
Reasoning for
Firecrawl is the research-scraping layer where an agent needs clean content, not raw HTML. In the Jakub Głąb Agent System it feeds the research/ingest skills — pulling articles and pages into the Markdown entity-wiki the agent reasons over — and it’s part of the 200IQ Labs shared-skills toolkit. The reason it earns a slot alongside Perplexity: Perplexity answers questions with citations, while Firecrawl gives you the full source content in a form an LLM can chew on — complementary halves of a research pipeline. This repo already ships a Firecrawl skill; this note is the corresponding tool reference.
Alternatives considered
- Scrapling — adaptive Python scraping framework already in the vault; more control/self-hosted, Firecrawl is the managed “just give me Markdown” option.
- Browser Use — agentic browser automation for interactive sites; heavier than Firecrawl for plain content extraction.
- Bright Data — industrial-scale scraping + proxies + datasets; Firecrawl is lighter and LLM-output-shaped.
- Jina Reader — similar URL→Markdown service; Firecrawl adds crawl/extract/search modes.
Resources
- 📘 Firecrawl docs
- 🧩 Extract (structured)
- 🔗 Perplexity · Scrapling · Bright Data — related research tooling
- Extending Claude Code — Tools for Its Blind Spots — compiled article featuring this tool
Template: tool