Website Screenshot & PDF API

Turn a list of URLs into full-page screenshots (PNG or JPEG) or PDFs in one run. Cookie banners are hidden, lazy images are loaded, and you get desktop, laptop, tablet or mobile views. You pay only for screenshots that succeed. Failed or timed-out pages cost nothing.

Built for AI agents and automations. Each result has a public image URL, the page title, the HTTP status, and optionally the page's visible text, so an agent can see and read a page in one call.

What you get

{
  "url": "https://example.com/pricing",
  "finalUrl": "https://example.com/pricing",
  "ok": true,
  "statusCode": 200,
  "title": "Pricing | Example",
  "screenshotUrl": "https://api.apify.com/v2/key-value-stores/<storeId>/records/0-example.com-1a2b3c4d",
  "format": "png",
  "device": "desktop",
  "width": 1920,
  "height": 4210,
  "fullPage": true,
  "truncated": false,
  "bytes": 912345,
  "text": "Plans for every team ... (only when includeText is on)",
  "capturedAt": "2026-10-18T14:03:11.000Z"
}

A failed page returns "ok": false with an error message. It is not charged.

Use cases

• AI agents: visual checks of a page, UI review, "what does this site look like right now".
• Monitoring: daily full-page snapshots of competitor pricing or landing pages, run on a schedule.
• QA: check a design on desktop, tablet and mobile in one batch.
• Archiving and compliance: PDF or PNG proof of what a page showed and when.
• Reports and decks: clean screenshots with no cookie popups.

Features

Full page or first screen	Long pages are cut at a height you choose (default 15,000 px) and flagged truncated
Devices	desktop 1920×1080, laptop 1366×768, tablet 820×1180 @2x, mobile 390×844 @3x (iPhone user agent), or a custom width and height
Formats	PNG, JPEG (quality setting), PDF (A4, with backgrounds)
Clean shots	Hides OneTrust, Cookiebot, Didomi, Quantcast, Usercentrics, TrustArc, Osano, CookieYes, Complianz and other consent popups without clicking accept. Add your own selectors to hide chat widgets and the like
Hard pages	Wait for a CSS selector, choose the load event (load, domcontentloaded, networkidle), add a delay, scroll to load lazy images, dark mode
Element shots	Capture only one element, e.g. #pricing
Text for LLMs	includeText returns the page's visible text, capped at maxTextChars
Batch	Runs pages in parallel, retries failed pages, and works with an optional Apify Proxy
Budget-safe	Respects your max cost per run and stops before going over

Pricing

Pay per event:

• Screenshot: about $3 per 1,000 successful screenshots or PDFs. Volume tiers are cheaper.
• Actor start: a small fee per run, which covers starting the browser.

There is no charge for pages that fail. Platform usage is included.

Input example

{
  "urls": ["https://apify.com", "news.ycombinator.com"],
  "device": "mobile",
  "fullPage": true,
  "format": "png",
  "hideCookieBanners": true,
  "includeText": true
}

Using it from an AI agent (MCP)

This actor works with the Apify MCP server. Agents can find it with the search tool and call it with urls. The result's screenshotUrl is a direct image link that vision models can open. It uses pay-per-event pricing without Standby, so agents can also pay per call where Apify supports agent payments.

Tips

• If a site loads content late, set waitUntil: "networkidle" or a waitForSelector.
• For sites that block data-center traffic, turn on a proxy in Proxy.
• For a large batch, raise Parallel pages and run with 4 GB of memory.

Limits and fair use

The actor captures public pages that need no login. Only capture pages you are allowed to access, and follow each site's terms. Images are stored in the run's key-value store and follow your Apify data retention settings.

Feedback

Found a page that renders badly? Open an issue on the Issues tab with the URL and your input.