Website Screenshot & PDF: Full Page, Devices, No Cookie Banners

Take clean screenshots of any public web page in bulk: PNG, JPEG, WebP or PDF, full page or viewport, on desktop, laptop, tablet or mobile. Cookie and consent banners are hidden without accepting anything, lazy-loaded images are loaded by scrolling through the page first, and every file comes back as a public link.

What you get per URL

field	meaning
url, finalUrl	the requested URL and where it ended up after redirects
httpStatus, title	HTTP status of the page and its title
device, viewport	preset used, with width, height and pixel density
files	one entry per format: format, key, url (public link), bytes, width, height (pixels; for PDF the page count is in pages)
bannerHidden	what was removed, e.g. Sourcepoint, OneTrust or overlay: div#cmp-banner
loadMs	time until the page was loaded
viaProxy	whether the page was loaded through the proxy
error	why a URL failed (invalid URL, DNS, timeout, HTTP 404, selector not found)

File keys are stable and readable: host, path, device and format, e.g. www.spiegel.de-desktop.png or www.spiegel.de-politik-mobile.jpg. The same URL always gets the same key, which makes comparing runs easy.

Features

• Formats: PNG, JPEG, WebP and PDF from one page load. JPEG and WebP quality adjustable.
• Devices: desktop 1440x900, laptop 1280x800, tablet 820x1180, mobile 390x844 with mobile user agent, touch and 3x pixel density, or any custom size.
• Element capture: pass a CSS selector to capture one element only.
• Wait strategy: load event or network idle, plus an extra delay.
• Lazy images: scrolls through the page before capturing.
• Cookie banners: known consent platforms (OneTrust, Cookiebot, Usercentrics, Sourcepoint, Didomi, Quantcast, Borlabs, Complianz, Klaro, CookieYes, consentmanager, Iubenda, TrustArc and more) plus a heuristic for large fixed overlays that mention cookies or consent. The banner is removed from the page, never clicked, so no consent is given.
• Dark mode, hide elements (chat widgets, ads), PDF in A4 or Letter, portrait or landscape, with or without backgrounds.
• Reliable: one browser with an isolated context per URL, a hard timeout per URL, automatic browser restart after a crash, one retry through the proxy when a page blocks.

Pricing

You pay per URL with at least one file delivered, no matter how many formats you pick. Failed URLs (invalid, unreachable, timeout, HTTP errors) are not charged.

Input example

{ "urls": ["https://www.spiegel.de/"], "formats": ["png", "pdf"], "device": "mobile", "fullPage": true }

Limits

• Only public pages. No logins, no CAPTCHA solving.
• Full-page images are cut at 16,000 CSS px height by default (adjustable up to 30,000 px). Long news front pages on the mobile preset (3x density) can produce PNGs above 20 MB; pick JPEG or WebP (about 5 to 10 times smaller) or set the pixel density to 1.
• The banner heuristic is careful on purpose: an unusual banner may stay visible. Add its selector to hideSelectors.
• Some sites block datacenter traffic. With the automatic proxy strategy those pages are retried through the proxy with a realistic browser fingerprint.
• Shops and publishers with strong bot protection can still block every automated browser. Such URLs, and pages that render empty, end as an error item and are not charged.
• With selector, PDFs still contain the whole page.

FAQ

What am I charged for? One event per URL with at least one file delivered, no matter how many formats you pick. Items with an error are not charged.

Where are the files stored? In the run's default key-value store under a stable key built from host, path, device and format. Each entry in files contains the key and a public url to the record.

Do I need a proxy? Not usually. With the default proxyStrategy auto pages load directly, and a page that answers 403, 429 or 503, shows a bot check or stalls is retried once through the proxy with a realistic browser fingerprint. direct never uses a proxy, proxy always does.

Are cookie banners accepted? No. Known consent platforms and large fixed cookie overlays are removed from the page, never clicked, so no consent is given. What was removed is listed in bannerHidden.

What happens when a page fails? The URL still gets a dataset item with error set, for example for an invalid URL, a timeout (default 90 seconds per URL, timeoutSecs 15 to 300), an HTTP error page or a page that renders empty because of bot protection.

Use with AI agents (MCP)

This Actor works as a tool for Claude, ChatGPT, Cursor and other MCP clients through the Apify MCP server. Add only this Actor:

https://mcp.apify.com?tools=webify-lab/website-screenshot-pdf

On first use the server signs you in with OAuth. Alternatively send the header Authorization: Bearer <APIFY_TOKEN>. Agents on the default Apify MCP setup can also find this Actor with search-actors and run it with call-actor. Pricing is the same as in Apify Console: you pay per result.