Website Screenshot API — Full Page, Mobile, No Cookie Banners

Capture clean screenshots of public web pages in bulk: the whole page or just the first screen, on desktop, tablet or mobile — with lazy-loaded images actually loaded, sticky headers not repeated, and cookie banners and sign-in pop-ups out of the way.

Failed screenshots are never charged. Pages that return an error, refuse automated access, show a bot check, time out or render empty are reported in the dataset with charged: false. Each run also has a small start fee (see Pricing), which applies even if every page fails.

Who it is for

• SEO and marketing agencies — weekly full-page snapshots of client and competitor landing pages.
• No-code automations (Make, n8n, Zapier) — "take a screenshot of this URL" as one reliable step.
• AI agents and QA — a visual of what a page really looks like, on desktop and mobile.

Example: an agency monitors 40 competitor pricing pages every Monday. It runs this Actor on a schedule with the 40 URLs in fullPage mode and gets 40 PNGs plus a table of which pages failed — failed pages are not billed.

What it fixes compared with a plain screenshot

Problem	What this Actor does
Images below the fold are blank	scrolls the page to trigger lazy loading, then returns to the top
Sticky header repeated / covering content in full-page shots	turns fixed and sticky elements into normal page flow before capture
Cookie banner covers the page	hides known consent managers and consent overlays — nothing is accepted on your behalf
Sign-in / newsletter pop-ups	hides modal dialogs, dark backdrops and floating sign-in widgets (hidePopups)
Can't choose the height	viewport, fullPage (up to 20,000 px) or fixedHeight (top N px)
Emoji show as boxes	colour-emoji and CJK fonts are installed in the image
Stuck runs	hard timeout per URL (urlTimeoutSecs), failed pages are not charged

Pricing

Pay per event. Prices are set in Apify Console; this table explains them.

Event	Price	When
Full-page screenshot	$0.004	per page captured successfully in fullPage or fixedHeight mode
First-screen screenshot	$0.002	per page captured successfully in viewport mode
Actor start	$0.001 per GB of memory	once per run, charged by Apify at start: $0.004 with the default 4 GB, $0.002 with 2 GB

Examples with the default 4 GB: 1 first-screen shot = $0.006; 40 full-page shots = $0.164; a run where every page fails = $0.004. Viewport-only runs work with 2 GB of memory (set it in the run options) and pay a $0.002 start fee. No separate platform-usage charge is added on top of these events.

Input

{
  "urls": ["https://www.bbc.com/news", "https://stripe.com", "https://en.wikipedia.org/wiki/Emoji"],
  "mode": "fullPage",
  "device": "desktop",
  "format": "png",
  "hideCookieBanners": true,
  "hidePopups": true,
  "blockAds": true
}

Field	Default	Notes
urls	—	1–500 public http(s) URLs; duplicates are captured once; [{"url": ...}] objects also accepted
mode	fullPage	viewport, fullPage (cut at 20,000 px), fixedHeight
height	1080	CSS pixels, used with fixedHeight
device	desktop	desktop 1280 px, tablet 820 px (2× density), mobile 390 px (3× density)
format	png	png, jpeg, webp; JPEG/WebP files are 3–10× smaller
quality	85	JPEG/WebP only
waitMs	0	extra wait after load (animations)
hideCookieBanners, hidePopups, blockAds	true	see table above
hideSelectors	[]	your own CSS selectors to hide (chat widgets, promo bars)
urlTimeoutSecs	90	per URL (5–300); slower pages are reported as failed and not charged

Image width/height in the output are in device pixels (a 390 px wide mobile viewport at 3× density gives a 1170 px wide image).

Output

One dataset item per URL; images are stored in the run's key-value store and linked by screenshotUrl.

{
  "url": "https://www.bbc.com/news",
  "finalUrl": "https://www.bbc.com/news",
  "httpStatus": 200,
  "status": "ok",
  "screenshotUrl": "https://api.apify.com/v2/key-value-stores/<store>/records/screenshot-0001-….png",
  "width": 1280, "height": 6541, "bytes": 3589541,
  "title": "BBC News - Breaking news, video and the latest top stories …",
  "error": null, "mode": "fullPage", "device": "desktop", "format": "png",
  "durationMs": 4135, "capturedAt": "2026-10-06T10:02:11+00:00", "charged": true
}

status is ok, blocked (the site refused automated access, HTTP 401/403/429/451 or a bot check — never bypassed) or error (DNS/TLS/timeout/HTTP error/empty page/invalid URL). Only ok items are charged.

Limits and honest notes

• Public pages only: no logins, no cookies/sessions, no proxies, no CAPTCHA solving. Sites that block automated browsers return blocked.
• Cookie banners are hidden visually; the page is captured as a first-time visitor who has not consented, so some sites show reduced content.
• Very long pages are cut at 20,000 px; video is not loaded (poster frames only) when blockAds is on.
• Very long documents are cut at 20,000 px before rendering, so even 100,000-px pages are captured in seconds.
• Emoji added in the newest Unicode release (16.0) may still show as boxes until the system fonts catch up.
• Default memory is 4 GB (heavy pages need more than 2 GB); for viewport-only runs 2 GB is enough.
• You are responsible for having the right to capture and use screenshots of the pages you submit.

Run locally

pip install -r requirements.txt && playwright install chromium
mkdir -p storage/key_value_stores/default && cp example_input.json storage/key_value_stores/default/INPUT.json
python -m src