Complete Page Capture - Screenshot & PDF

Capture complete full-page screenshots or PDFs of any URL list. This actor scrolls lazy content into view, waits for images and fonts, removes fixed headers and consent overlays, supports real device emulation, and tells you when a page was blocked or truncated instead of returning a broken image as a success.

Why this actor

• Complete captures - scrolls through the page to trigger lazy images and infinite-scroll widgets, then captures with the page reset to the top.
• Clean captures - can hide fixed headers, overlays and common cookie banners before the shot without deleting sticky content cards.
• Honest results - reports HTTP status, page title, block reason, truncation, image failures and file metadata for every URL.
• Real device modes - desktop, laptop, tablet, iPhone, Pixel and Galaxy presets with the right viewport, user agent, touch support and pixel density.
• Useful formats - PNG, JPEG, WebP and PDF, with PDF paper size, orientation, margins, backgrounds and a one-page mode.
• Bulk friendly - concurrent captures, bounded retries, per-URL timeouts and a hard cap for very long or infinite pages.

Quick start

{
  "urls": [
    {"url": "https://www.apify.com/"},
    {"url": "https://en.wikipedia.org/wiki/World_War_II"}
  ],
  "format": "png",
  "device": "desktop",
  "scrollToBottom": true
}

Every URL produces one dataset record and one file in the run's key-value store. The screenshotUrl field is a public download link; no Apify token is needed to open it.

Inputs

Field	Default	What it does
urls	-	Required. URL list; accepts URL objects or plain strings. Bare domains use https.
format	png	png, jpeg, webp or pdf.
quality	85	JPEG/WebP quality, 1-100.
fullPage	true	Capture the whole scrollable page. Turn off for the first screen only.
selector	-	Capture one element by CSS selector instead of the page.
device	custom	custom, desktop, desktop_hd, laptop, tablet, ipad, iphone_15, iphone_se, pixel_8, galaxy_s23.
viewportWidth	1280	Width for device=custom, 100-3840.
viewportHeight	1080	Height for device=custom, 100-4320.
deviceScaleFactor	1	Pixel density for device=custom, 1-3.
colorScheme	no-preference	no-preference, light or dark.
locale	en-US	Browser language for localized pages.
waitUntil	load	load, domcontentloaded, networkidle or commit.
waitForSelector	-	Wait for a selector to become visible before capture.
waitForImages	true	Wait up to 5 seconds for images after scrolling.
waitForFonts	true	Wait up to 5 seconds for web fonts.
delay	0	Extra milliseconds after navigation.
scrollToBottom	true	Scroll through the page to trigger lazy content.
maxScrollSteps	400	Hard scroll-step cap.
scrollStepPx	250	Pixels per scroll step.
delayAfterScrolling	1000	Milliseconds to settle after scrolling.
waitUntilNetworkIdleAfterScroll	false	Wait for network idle after scrolling.
waitUntilNetworkIdleAfterScrollTimeout	15000	Network-idle timeout in milliseconds.
hideFixedElements	false	Hide fixed headers, full-screen overlays and chat widgets. Sticky content blocks stay in place.
cookieBannerAction	hide	hide, accept or off.
selectorsToHide	-	Extra comma/newline-separated CSS selectors to hide.
pdfPageSize	web	web, a4, letter, legal, a3, a5 or tabloid.
pdfLandscape	false	Landscape PDF pages.
pdfPrintBackground	true	Include CSS backgrounds and colors in the PDF.
pdfFitOnePage	false	One tall PDF page instead of paginated paper.
pdfMargin	none	none, narrow, normal or wide.
maxConcurrency	1	URLs captured in parallel, 1-5. Use 1 for long pages at 2048 MB.
maxRetries	1	Retries for transient navigation failures and blocks.
navigationTimeoutSecs	60	Per-URL navigation and selector timeout.
maxPageHeight	0	Clip full-page images at this CSS height; 0 is unlimited. PDFs keep their full document height.
proxyConfiguration	-	Optional proxy override. The actor starts direct and escalates to an AUTO datacenter proxy only after a real block.

Output

Each dataset record contains:

Field	Meaning
startUrl	URL you submitted.
url	Final URL after redirects.
screenshotUrl	Public download link for the file.
screenshotKey	Key of the file in the run's key-value store.
format	png, jpeg, webp or pdf.
status	success for a real stored capture, otherwise error: ....
timestamp	ISO-8601 UTC capture time.
pageTitle	Page title at capture time.
httpStatus	Main document HTTP status.
contentType	Main response content type.
blocked / blockReason	Whether the page was a block, captcha or login wall, and which kind.
errorType	Machine-readable error category.
device, viewportWidth, viewportHeight, deviceScaleFactor, colorScheme	Rendering settings used.
fullPage, selector	Capture scope.
imageWidth, imageHeight, fileSizeBytes	File dimensions and size.
pageHeight, capturedHeight, truncated	Page height, captured height and whether a cap clipped it.
pdfPages, pdfPageSize	PDF page count and paper size.
durationMs, imagesTotal, imagesFailed, retries	Capture evidence. Image counts cover visible images only; hidden tracking pixels are ignored.
warnings	Human-readable notes such as failed images or a hit scroll cap.

Blocked pages are not stored as screenshots and do not produce a success record. The record says blocked: true and status: error: blocked (...), so a challenge page cannot be mistaken for page content.

Examples

Mobile screenshot

{
  "urls": [{"url": "https://stripe.com/"}],
  "device": "iphone_15",
  "format": "jpeg",
  "quality": 85
}

Full-page, dark PDF

{
  "urls": [{"url": "https://tailwindcss.com/"}],
  "format": "pdf",
  "colorScheme": "dark",
  "pdfPageSize": "a4",
  "pdfFitOnePage": true,
  "cookieBannerAction": "hide"
}

Capture one section

{
  "urls": [{"url": "https://www.apify.com/"}],
  "selector": "main",
  "fullPage": false,
  "format": "png"
}

What changed in this revamp

• scrollToBottom now defaults to true so lazy images are actually loaded.
• hideFixedElements defaults to false: resetting to the top already avoids repeated headers on most pages, and this keeps sticky content cards in the image. Turn it on to remove fixed headers/overlays.
• cookieBannerAction now defaults to hide.
• maxConcurrency now defaults to 1 for safe 2048 MB memory use; raise it only when you have memory headroom.
• The screenshotUrl field now returns a working public link.
• waitUntil=domcontentloaded is the prefill because networkidle can stall on analytics-heavy pages.
• Pages that return 403/captcha/login walls are reported as blocked errors, not successes.

You can turn each of these off from the input form.

Limitations

• Only public, anonymous pages are supported. Login-protected pages are outside this actor's design.
• Some sites use CAPTCHA or bot defenses that cannot be bypassed. Those URLs are reported as blocked and are not charged as successful captures.
• Consent banners are handled on a best-effort basis. Hide is more reliable than click-accept; use selectorsToHide for a site-specific banner.
• Some lazy images never load because of the site's own rules or missing assets. Only visible images are counted in imagesFailed and listed in warnings; hidden tracking pixels and empty placeholders are ignored.
• Fixed headers are kept by default. Enable hideFixedElements to remove fixed headers, full-screen overlays and chat widgets; sticky content cards are intentionally preserved.
• Very long or infinite pages are capped by maxScrollSteps and maxPageHeight; the record sets truncated: true when a cap clips the capture.
• A direct PDF URL is copied when format=pdf; asking for an image of a PDF is reported as an error because Chromium downloads the file instead of rendering it.
• PDF pagination follows the browser's print layout, which can differ from the screen layout.

FAQ

Do I need a proxy? No. The actor starts with a direct connection and only uses an AUTO datacenter proxy after a real block. If you supply your own proxy, it is used instead.

Can I capture a page that requires a login? No. Authenticated pages are intentionally out of scope.

How do I hide a cookie banner? Leave cookieBannerAction on hide. If a site uses a custom banner, add its selector to selectorsToHide.

Why is a page reported as blocked? The actor saw an HTTP 403/429/451, a CAPTCHA/Cloudflare challenge, or a login wall. The block page is not saved as a normal screenshot.

Why is my image shorter than the page? Check truncated and warnings. maxPageHeight or maxScrollSteps may have capped the capture.

Can I capture multiple URLs? Yes. Each URL produces its own record and file. Keep maxConcurrency at 1 for very long pages at 2048 MB; use a larger memory setting before raising concurrency.

Can I make a single-page PDF? Set format=pdf and pdfFitOnePage=true. For paper output, choose pdfPageSize and pdfMargin.