Firecrawl
Connects to the Firecrawl web-scraping API. An agent can scrape a single page to clean markdown, run a web search (optionally scraping the results), or kick off an asynchronous crawl of an entire site.
Add to an agent node's tools array with _node_type: "tool-firecrawl". See Connector Reference — How connectors are used.
Authentication
| Field | Required | Description |
|---|---|---|
api_key | Yes | Firecrawl API key, sent as Authorization: Bearer <api_key>. Typically a stored secret. |
{
"name": "firecrawl",
"_node_type": "tool-firecrawl",
"api_key": "{{ secret.FIRECRAWL_API_KEY }}"
}
Tools
firecrawl_scrape
Scrape a single web page and return its content as clean markdown. Great for reading documentation, articles, and web pages.
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
url | string | Yes | URL to scrape |
formats | string | No | Output formats: markdown,html,links (default markdown) |
Returns: The page content as markdown, or "No content extracted." if Firecrawl returned no markdown.
firecrawl_search
Search the web and optionally scrape the content of result pages using Firecrawl.
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
query | string | Yes | Search query |
limit | number | No | Number of results (default 5) |
scrape | boolean | No | Also scrape the page content (default false) |
Returns: A list of results, each with title, URL, and description (N. **title**\n url\n description); if scrape is true, each result also includes its scraped markdown content. Returns "No results found." if there are no matches.
firecrawl_crawl
Crawl a website and return content from multiple pages as markdown. Use for indexing docs or entire sites.
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
url | string | Yes | Starting URL to crawl |
max_pages | number | No | Max pages to crawl (default 10) |
max_depth | number | No | Max crawl depth (default 2) |
Returns: A confirmation message with the started crawl job's ID and the status-check endpoint (Crawl job started. Job ID: <id>\nCheck status at: GET https://api.firecrawl.dev/v1/crawl/<id>). The crawl runs asynchronously on Firecrawl's side — this tool does not poll for completion.