Firecrawl Block

The Firecrawl block extracts clean, structured content from any website with advanced web scraping, crawls entire sites, or searches the web for information. Reach for it whenever a workflow needs page content, multi-page crawls, or live search results to feed downstream blocks.

Overview

When to Use

• Scrape a single web page and get clean markdown, HTML, and metadata
• Focus extraction on the main content of a page, stripping nav and boilerplate
• Crawl an entire website up to a page limit to gather many pages at once
• Search the web for a query and retrieve structured result data
• Feed scraped or searched content into an Agent or other downstream block
• Monitor or ingest external sites as part of an automated pipeline

Configuration

Operation

Dropdown selecting what Firecrawl does. Defaults to scrape. The remaining fields are shown based on this choice.

Website URL

Required for scrape and crawl operations. The target page or site URL to extract content from.

Only Main Content

Switch shown for the scrape operation only. When enabled, returns just the primary content of the page, dropping navigation, footers, and other boilerplate.

Page Limit

Short input shown for the crawl operation only. Maximum number of pages to crawl. Defaults to 100 when not specified.

Search Query

Required for the search operation. The query terms used to search the web.

API Key

Required for all operations. Your Firecrawl API key (stored securely as a password field).

Inputs & Outputs

• Inputs:

  • apiKey (string) — Firecrawl API key
  • operation (string) — Operation to perform (scrape, search, or crawl)
  • url (string) — Target website URL (used by scrape and crawl)
  • limit (string) — Page crawl limit (used by crawl)
  • query (string) — Search query terms (used by search)
  • scrapeOptions (json) — Scraping options (e.g. custom formats; hidden from UI)

• Outputs:

  • markdown (string) — Page content in markdown format (scrape)
  • html (string) — Raw HTML content (scrape)
  • metadata (json) — Page metadata including SEO and Open Graph information (scrape)
  • data (json) — Search results data; each item contains title, description, url, markdown, html, rawHtml, links, screenshot, and metadata (search)
  • warning (string) — Warning messages from the operation (search)
  • pages (json) — Array of crawled pages, each with markdown, html, and metadata (crawl)
  • total (number) — Total number of pages found during crawl (crawl)
  • creditsUsed (number) — Number of credits consumed by the crawl operation (crawl)

Tools

Firecrawl Website Scraper (firecrawl_scrape) — Extracts structured content from a single web page. Converts the page to markdown and/or HTML while capturing SEO metadata, Open Graph tags, and status information. Uses params: url, apiKey, and optionally scrapeOptions (formats array).

Firecrawl Search (firecrawl_search) — Searches the web using Firecrawl and returns an array of structured result objects (title, description, url, markdown, html, rawHtml, links, screenshot, metadata). Uses params: query and apiKey.

Firecrawl Crawl (firecrawl_crawl) — Crawls an entire website asynchronously and extracts structured content from all accessible pages up to a configurable limit. Submits a crawl job and polls for completion (up to 5 minutes). Uses params: url, apiKey, limit (max pages, default 100), and onlyMainContent (boolean).

YAML Example

firecrawl_1:
  type: firecrawl
  name: "Firecrawl"
  inputs:
    operation: "scrape"
    url: "{{start.input}}"
    apiKey: "{{FIRECRAWL_API_KEY}}"
  connections:
    outgoing:
      - target: next-block-id