🎨 Enhanced Visual Asset Extraction Suite

Extract all visual assets from any website with intelligent categorization and rich metadata, built as a Python Apify Actor for e‑commerce, design, and automation workflows.

---

🚀 What this actor does

This actor crawls a list of website URLs and extracts visual assets such as:

• 🏢 Logos & brand marks (including favicons)
• 🛍️ Product images and catalog thumbnails
• 🎯 Hero / banner images (headers, mastheads)
• 👥 Team photos, avatars, profile pictures
• 📱 Social media graphics (Open Graph, Twitter cards)
• 🎨 Icon sets (SVG, PNG, touch icons)
• 🖼️ Gallery / slider / carousel images

For each asset, it attempts to infer:

• Category (logo, product, hero, team, social, icon, gallery, other)
• Source (HTML tag / location where it was found)
• Format (SVG/PNG/JPG/WebP/AVIF/ICO/…)
• Optional dimensions (width, height) when enabled

---

🧠 How it works

The actor:

1. Downloads the HTML of each URL.
2. Parses the page and extracts images from:
   • <img> tags (including lazy‑loading attributes like data-src)
   • <meta property="og:image"> and name="twitter:image"
   • <link rel="icon">, Apple touch icons, and default /favicon.ico
   • <picture> elements and srcset attributes
   • Inline CSS background-image: url(...) styles
3. Converts relative URLs to absolute.
4. Applies a rule‑based classifier that looks at:
   • Image URL
   • alt text
   • CSS classes
   • Local DOM context
     to categorize each asset.
5. Optionally fetches images and uses Pillow to read true width, height, and format.
6. De‑duplicates assets by URL and outputs a structured JSON object per input URL.

---

⚙️ Input

The actor accepts a JSON object with the following fields:

{
  "urls": [
    "https://www.apple.com",
    "https://www.nike.com"
  ],
  "extractDimensions": true,
  "fetchDimensions": false,
  "maxConcurrency": 5
}


Input fields
urls (array of strings, required)
List of website URLs to extract visual assets from.

extractDimensions (boolean, default: true)
If true, detects image format (SVG/PNG/WebP/etc.) from the URL and enriches output with this information.

fetchDimensions (boolean, default: false)
If true, downloads each image (under 5 MB) and uses Pillow to read actual width and height.
This is slower and uses more resources, but yields precise dimensions.

maxConcurrency (integer, default: 5, min: 1, max: 20)
Maximum number of URLs processed concurrently.

{
  "url": "https://example.com",
  "totalAssets": 47,
  "categoryBreakdown": {
    "logo": 3,
    "product": 24,
    "hero": 2,
    "team": 8,
    "icon": 6,
    "social": 2,
    "other": 2
  },
  "assets": [
    {
      "src": "https://example.com/images/logo.svg",
      "category": "logo",
      "format": "svg",
      "source": "img_tag",
      "alt": "Company Logo",
      "class": "site-logo",
      "width": "200",
      "height": "60",
      "loading": "lazy"
    },
    {
      "src": "https://example.com/images/hero-banner.webp",
      "category": "hero",
      "format": "webp",
      "source": "css_background",
      "class": "homepage-hero",
      "width": 1920,
      "height": 1080
    }
  ]
}