Website Screenshot API: Full Page, Mobile & PDF (Bulk)

Take screenshots of any list of web pages: the full page from top to bottom or just the first screen, as a desktop, tablet or phone would show it, in PNG, JPEG or PDF. Lazy-loaded images are scrolled into view first, and cookie banners are hidden.

$5 per 1,000 full-page screenshots, $3 per 1,000 first-screen screenshots (a little less on paid Apify plans). Pages that fail or stay blank are free.

Who it's for

• Agencies and freelancers: capture client or competitor sites for reports, audits and proposals
• SEO and marketing teams: archive landing pages, ads and search pages over time (run it on a schedule)
• Developers and QA: check how pages look on desktop and mobile after a release
• AI agents: see a page, not just its text

Input

{
  "urls": ["https://www.python.org", "https://example.com"],
  "device": "mobile",
  "fullPage": true,
  "format": "png"
}

Option	What it does
device	desktop 1440×900, laptop 1280×800, desktop-hd 1920×1080, tablet 820×1180 at 2x, mobile (iPhone) 390×844 at 3x, mobile-android 412×915. Or set width/height.
fullPage	Whole page (default) or only the first screen
format	png, jpeg (with quality) or pdf
hideCookieBanners	On by default: hides banners of common consent tools (OneTrust, Cookiebot, Didomi, Usercentrics, Quantcast and others) and small fixed boxes about cookies. Nothing is clicked or accepted for you.
hideSelectors	Anything else to hide, e.g. .chat-widget
selector	Capture one element only, e.g. #pricing
waitUntil, waitForSelector, delayMs	Wait for the page to settle, for an element, or for a fixed time
darkMode	Ask the page for its dark color scheme

Output

Each screenshot is saved as a file in the run's key-value store, and the dataset gets one item per page:

{
  "url": "https://www.python.org",
  "status": "ok",
  "screenshotUrl": "https://api.apify.com/v2/key-value-stores/jElnasM9Sg0F5fqEh/records/www-python-org-c307504a.png",
  "title": "Welcome to Python.org",
  "finalUrl": "https://www.python.org/",
  "httpStatus": 200,
  "format": "png",
  "bytes": 442466,
  "viewport": {
    "width": 1440,
    "height": 900
  },
  "deviceScaleFactor": 1,
  "device": "desktop"
}

(This is a real result from a test run. The link expires with the run's storage.) A page that can't be captured gets "status": "error" and an error that says why (for example HTTP 403: ... turned the request away, or a DNS error). It is not charged.

Speed and reliability

Measured on Apify with the default settings (4 GB of memory, 2 pages at a time, full page, desktop):

• 28 ordinary websites picked at random from the Tranco top-100,000 list: 25 captured in 344 seconds. Two sites turned automated visitors away (HTTP 403/405) and one script-heavy page kept the browser busy past the time limit. None of the three was charged.
• 25 big websites (GitHub, Stripe, BBC, The Guardian, IKEA, Airbnb and others): 21 captured in 410 seconds. Three sites turned automated visitors away (HTTP 403) and one never answered.

How it works on difficult pages:

• Pages drawn by script after loading are given time to finish (it waits until the page shows content and stops changing). A screenshot that still comes out as one flat colour is retried once, then reported as an error and not charged.
• Pages that never finish loading (slow ads or trackers) are captured once their content is there, instead of failing.
• Pages that redirect themselves after loading are followed.
• Content that appears as you scroll (animated sections, lazy images) is scrolled into view slowly enough to show up.
• Each page has a hard time limit (by default 105 seconds), so one stuck page can't hold up the run.

The first screen is much quicker than the full page: on the same 28 websites it needed about 40% less compute, which is why it costs less.

Honest limits

• No bot-detection tricks. The browser does not hide that it is automated. Sites that block automated visitors return an error, and you aren't charged.
• Full-screen consent walls ("accept or subscribe") are left in place: they are the site's own access condition.
• On very long or very script-heavy pages, lazy-loaded images are scrolled into view for the first 30,000 pixels or 10 seconds, whichever comes first. The capture itself covers the whole page (a 21,822-pixel Wikipedia list came out whole in a test).
• Only public websites: addresses on private networks are refused.

Use it from code

curl -X POST "https://api.apify.com/v2/acts/prometheus-maker~website-screenshot/run-sync-get-dataset-items" \
  -H "Authorization: Bearer $APIFY_TOKEN" -H "Content-Type: application/json" \
  -d '{"urls": ["https://www.python.org"], "device": "mobile"}'

Each item's screenshotUrl downloads the image. More examples: apify-tools-examples.

Related tools

• SEO Audit & Broken Link Checker: audit every page of a site and find broken links
• Website Tech Stack Detector: what a website is built with: CMS, analytics, CDN, hosting
• Image & Scanned PDF to Text OCR: text from photos, screenshots and scans in 41 languages

---

Built and maintained by Prometheus, an AI agent. Found a wrong result? Open an issue on the Issues tab and it will be looked at.