Website Image Scraper

Extract every image URL from a website. Crawls the start page (and optionally internal links up to a configurable depth), then parses <img> tags, <picture>/<source>, srcset candidates, <link rel="icon">, and CSS background-image declarations. HTTP-only — no browser, no proxy, no API key.

What it does

• Pull every image URL referenced on a page — <img src>, lazy-loaded data-src, srcset candidates, picture sources, favicons, inline style="background-image: url(...)".
• Crawl deeper — follow internal links up to maxCrawlDepth (same host only) to grab images from linked pages too.
• Filter by format — restrict to specific extensions (e.g. only SVG, only WebP/AVIF).
• Bounded — maxImagesPerPage and maxTotalImages keep runs cost-predictable on large galleries.

Input

Field	Type	Default	Description
startUrl	string (required)	https://apify.com	Page to start crawling. Must be http:// or https://.
maxCrawlDepth	integer	1 (0–5)	0 = only the start URL; 1+ = follow internal links one level (same host only).
maxImagesPerPage	integer	200 (1–5000)	Cap per page — keeps pathological galleries bounded.
maxTotalImages	integer	1000 (1–50000)	Hard cap on total images emitted across the whole run.
imageExtensions	array	[jpg, jpeg, png, gif, webp, svg, avif, bmp, ico]	Only URLs whose path ends in one of these are kept.
includeBackgroundImages	boolean	true	Also extract from inline style="background-image: url(...)".
userAgent	string	(Chrome 131)	Optional UA override.

Example input

{
  "startUrl": "https://apify.com",
  "maxCrawlDepth": 1,
  "maxImagesPerPage": 200,
  "maxTotalImages": 500,
  "imageExtensions": ["jpg", "png", "webp", "svg"],
  "includeBackgroundImages": true
}

Output

One record per unique image URL. Empty fields are omitted (no nulls).

{
  "url": "https://apify.com/static/hero.jpg",
  "sourcePage": "https://apify.com/",
  "pageTitle": "Apify · The full-stack web-scraping & automation platform",
  "alt": "Apify hero image",
  "hasAltText": true,
  "title": "Apify",
  "width": 1200,
  "height": 600,
  "extension": "jpg",
  "discoveredVia": "img-tag",
  "mimeTypeHint": "image/jpeg",
  "crawlDepth": 0,
  "scrapedAt": "2024-12-16T14:23:11+00:00"
}

Output fields

• url — absolute URL of the image (data: URIs and javascript: pseudo-URLs are filtered out).
• sourcePage — the page where the image was discovered.
• pageTitle — <title> of the page where the image was found (handy for grouping the dataset by page name).
• alt — alt attribute of the <img> tag (when present).
• hasAltText — derived boolean: true when alt is present and non-empty. Lets you filter accessibility issues without testing for field presence.
• title — title attribute (when present).
• width / height — explicit pixel dimensions from the tag (only emitted when numeric).
• extension — lowercase file extension parsed from the URL path (e.g. "jpg", "svg", "webp"). Useful for format-bucket aggregations.
• discoveredVia — one of img-tag, srcset, picture-source, link-icon, css-background.
• mimeTypeHint — derived from the file extension (e.g. image/png, image/svg+xml).
• crawlDepth — depth at which the page was crawled (0 = startUrl).
• scrapedAt — ISO-8601 timestamp.

Use cases

• Content audits — see every image a website serves up, broken down by source (img tag vs CSS background).
• Asset inventory — pull all logos, hero images, and icons from a competitor or brand site.
• Format migration — find every JPEG/PNG to convert to WebP/AVIF, or every PNG to convert to SVG.
• SEO / accessibility — list images with hasAltText: false to flag accessibility issues at a glance.

FAQ

Does it download the image binaries? No. The actor only collects URLs and metadata. Combine with a separate downloader (or pipe URLs into Apify's standard "URL list" actor) if you need the bytes.

Does it work on JavaScript-rendered pages? Mostly no. This scraper is HTTP-only — it sees the server-rendered HTML, not what runs after the page boots. If a site lazy-loads images via React/Vue, you may only see fallback / placeholder images. For SPA-rendered content, use a Playwright-based actor instead.

Can I limit it to a single page? Set maxCrawlDepth: 0. Only the start URL is fetched.

Does it follow external links? No. Internal-link crawling only follows links to the same host as startUrl to keep cost and scope bounded.

What if the site has no images at all? You get a single sentinel record {"type": "website_image_scraper_error", "reason": "no_images_found"} so the dataset is non-empty. The run still completes successfully.

How does it deduplicate? By absolute URL. The same image referenced from multiple pages produces one record (the first-seen page is recorded as sourcePage).