Snapframe — Website Screenshots & PDFs

Capture pixel-perfect website screenshots (PNG/JPEG) and print-ready PDF documents in a real Chromium browser. Built by the team behind the Snapframe API.

Feed it one or more URLs and get back files plus a clean dataset row per page: direct download link, file size, page title, HTTP status and load time. Batch up to 50 URLs per run.

What it does

• Screenshots in PNG or JPEG — viewport or full-page, with device presets (desktop, laptop, tablet, mobile) or your own custom dimensions and scale factor.
• PDFs with paper size (A4/Letter/Legal), orientation, margins, background printing, and optional header/footer HTML templates with page numbers.
• First-class Arabic/RTL rendering — proper Arabic fonts ship in the image, so Arabic pages (invoices, articles, dashboards) render with real glyphs, not boxes.
• Batch captures — up to 50 URLs per run, each stored as its own file with a dataset record. One bad URL never kills the batch.
• Authenticated pages — pass custom request headers and cookies to capture pages behind a login.
• Safe by default — only public http(s) URLs are loaded; private, loopback and internal addresses are blocked.

Input

Field	Description
urls	URLs to capture (1–50, required)
outputFormat	png, jpeg or pdf
devicePreset	desktop (1920×1080), laptop, tablet, mobile, or custom
viewportWidth / viewportHeight / deviceScaleFactor	Used when the preset is custom
fullPage	Capture the whole scrollable page (images only; PDFs are always full)
waitUntil	load, domcontentloaded, networkidle or commit
delayMs	Extra pause after load for animations/lazy images
timeoutMs	Per-page load timeout
jpegQuality	JPEG quality 10–100
pdfFormat / pdfLandscape / pdfMargin / printBackground	PDF layout
displayHeaderFooter / headerTemplate / footerTemplate	PDF header/footer HTML (<span class="pageNumber"> supported)
headers	Extra HTTP headers, e.g. {"Authorization": "Bearer ..."}
cookies	Cookies injected before capture (each needs name, value and domain or url)

Example input:

{
  "urls": ["https://example.com", "https://news.ycombinator.com"],
  "outputFormat": "png",
  "devicePreset": "desktop",
  "fullPage": true,
  "waitUntil": "networkidle",
  "delayMs": 1000
}

PDF example:

{
  "urls": ["https://example.com"],
  "outputFormat": "pdf",
  "pdfFormat": "A4",
  "pdfMargin": "0.4in",
  "printBackground": true,
  "displayHeaderFooter": true,
  "footerTemplate": "<div style=\"font-size:8px;width:100%;text-align:center;\">Page <span class=\"pageNumber\"></span> of <span class=\"totalPages\"></span></div>"
}

Output

• Files — each capture is saved in the run's key-value store (0000-example-com.png, …) and stays downloadable via its direct URL.
• Dataset — one row per URL: url, ok, outputFormat, fileName, fileUrl, bytes, title, statusCode, loadTimeMs, and reason when a capture fails.
• OUTPUT — run summary: total / succeeded / failed counts and format.

Pricing

Pay per event — you are charged only for successful captures:

Event	Price
Screenshot captured (PNG/JPEG)	$0.003
PDF captured	$0.005

Platform compute costs are passed through to the run (standard Apify compute units), so the event price is the whole story. Failed or blocked URLs are never charged.

API examples

curl -X POST "https://api.apify.com/v2/acts/abdou_009~snapframe-apify-actor/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"urls": ["https://example.com"], "outputFormat": "pdf"}'

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("abdou_009/snapframe-apify-actor").call({
    "urls": ["https://example.com"],
    "outputFormat": "png",
    "devicePreset": "mobile",
})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["url"], "->", item["fileUrl"])

AI agents can also discover and call this actor through the Apify MCP server — the input schema is written to be agent-friendly.

Limits

• 50 URLs per run; schedule runs for larger batches.
• Pages behind logins need the headers/cookies inputs — there is no interactive login.
• Sites with aggressive bot protection may serve a challenge page; that is captured faithfully as what an anonymous visitor sees.
• Very long pages (tens of thousands of pixels) can exceed browser capture limits — use a tall custom viewport with fullPage: false for those.

Support

Issues: open one on the actor page. This actor is maintained by Snapframe — the same capture engine powers our hosted API.