Firecrawl
Firecrawl web scraping integration — turn websites into LLM-ready markdown, crawl entire sites, and extract structured data
The Firecrawl integration lets your agents scrape pages, crawl whole sites, generate sitemaps, and pull structured data out of arbitrary web pages — returning clean, LLM-ready markdown.
Prerequisites
- A Firecrawl API key from firecrawl.dev. Free tier works for evaluation; paid plans give higher rate limits and longer crawls.
Setup
- Open your agent in Agent Studio.
- In the Integrations panel, click Add Integration Config.
- Select Firecrawl from the integration list.
- Give the config a display name (e.g., "Firecrawl — Marketing scrape").
- Paste your API Key.
- Save the agent revision.
Using in Your Agent
- In an Agent Step, attach Firecrawl tools from the tools panel.
- Tools are referenced as
firecrawl.firecrawl_scrape,firecrawl.firecrawl_crawl, etc.
A scrape is one URL → markdown. A crawl walks the site graph and is more expensive; use it when you genuinely need multi-page coverage. map is the cheap "what URLs exist on this site" call.
Available Tools
| Tool | Description | Action Type |
|---|---|---|
firecrawl_scrape | Scrape a single page and return its content as clean markdown with configurable output formats | ReadOnly |
firecrawl_crawl | Crawl a website starting from a URL, following links to scrape multiple pages | ReadOnly |
firecrawl_map | Discover all URLs on a website by generating a sitemap without scraping content | ReadOnly |
firecrawl_extract | Extract structured data from a page using a natural language prompt or JSON schema | ReadOnly |