Actions

actions lets you run browser interactions (click/type/wait/scroll/execute JS) before extracting content.

Constraints

actions run sequentially
up to 50 actions per request
combined wait time (wait actions + waitFor) should not exceed 60s
actions are not supported for PDFs: if the URL resolves to a PDF, avoid actions or the request will fail

Wait by duration or wait until an element is visible (choose one).

{ type: 'wait', milliseconds: 1000 }
{ type: 'wait', selector: '#content' }

Field	Type	Notes
`milliseconds`	number	Fixed delay
`selector`	string	Wait until visible (30s timeout)

{ type: 'click', selector: '#accept' }
{ type: 'click', selector: '.expand', all: true }

Field	Type	Notes
`selector`	string	CSS selector
`all`	boolean	Click all matches

{ type: 'click', selector: '#q' }
{ type: 'write', text: 'firecrawl' }

Field	Type	Notes
`text`	string	Text to type (usually requires a prior click to focus)

{ type: 'press', key: 'Enter' }

Field	Type	Notes
`key`	string	e.g. Enter/Tab/Escape

{ type: 'scroll', direction: 'down' }
{ type: 'scroll', selector: '.list' }

Field	Type	Notes
`direction`	`"up" \| "down"`	Defaults to down
`selector`	string	Scroll a specific container

{ type: 'screenshot', fullPage: true, quality: 80 }

{ type: 'scrape' }

Captures current HTML into data.actions.scrapes.

{ type: 'executeJavascript', script: "document.title" }

Return values are captured in data.actions.javascriptReturns.

{ type: 'pdf', format: 'A4', landscape: false, scale: 1 }