Page Images Extractor: image URLs, alt text, sizes, lazy src

Every image on a web page, up to 500 pages a run and 1,000 images each: absolute URL, alt text, title, width and height, format, lazy source, srcset, whether it sits in a link, plus per page the counts of images and images without alt text. Robots-aware, no browser, no downloads. Pay per image.

Why this scraper

• Only delivered results are charged. Inputs that fail come back as clear error records at no cost.
• One plain fetch per page, parsed with lxml, no browser, no start fee and nothing downloaded. Measured on the platform: 32 images from three pages (a Wikipedia article, BBC News, python.org) in 6 seconds for a sixth of a cent.
• Every image as a usable row: the absolute URL (relative paths and <base> resolved), alt text with a separate missing-alt flag, title, width and height attributes, format from the URL, whether it is lazy-loaded and from which attribute, srcset candidate count and the largest candidate, whether the image sits inside a link and where that link goes, host and internal or external, and how often the page repeats it.
• Page totals on every row: unique image count, images without alt text and inline data: images, so an accessibility or SEO audit reads off the first row per page.
• Robots-aware: each host's robots.txt is read once per run and a closed path comes back as a free ROBOTS_DISALLOWED row (measured: Google's /search path).

Who this is for

Paste page URLs in urls (up to 500 per run) and set maxImagesPerPage (default 200, ceiling 1,000, page order). Switch includeDataUris on to see inline base64 images as rows with their length. Built for alt-text and accessibility audits, image SEO checks, finding oversized or lazy-loaded images, building image inventories of a site and seeding download or vision pipelines with clean URLs.

Who this is not for

No browser is used, so images that a page adds with JavaScript after loading are not seen; lazy-loading attributes in the HTML are read, scripts are not run. Only <img> elements count: CSS background images, <picture> sources without an <img>, videos and SVG drawn inline are not rows. width and height are the HTML attributes, not the pixel size of the file; format is read from the URL and empty when the URL has no extension or format parameter (measured: 1 of 32). Nothing is downloaded, so file sizes are not known. A page without any image comes back as a free NO_IMAGES row.

Input fields

Field	Type	Required or default	What it does
urls	list of text	required	One per row, up to 500: full page URLs (https://...). Each host's robots.txt is respected.
maxImagesPerPage	number	200	Cost ceiling per page, in page order.
includeDataUris	true/false	false	On: inline base64 images are delivered with their length instead of a URL. Off (default): they are counted but skipped.

Input example

{
    "urls": [
        "https://en.wikipedia.org/wiki/Web_scraping",
        "https://www.python.org/"
    ],
    "maxImagesPerPage": 200,
    "includeDataUris": false
}

Output example

Field	Type	What it holds	Example
pageUrl	text	The page this image was found on.	https://en.wikipedia.org/wiki/Web_scraping
position	number	The place in the source's own order, counting from 1.	1
url	text	The image file itself, made absolute.	https://en.wikipedia.org/static/images/icons/enwiki-25.svg
altText	text	The alt text, which is what a screen reader says and what search engines read. Empty is the finding, not a gap in our data.	Wikipedia
hasAlt	true/false	Whether the image has an alt attribute at all, which is not the same as having text in it.	False
title	text	Filled when the image carries a title attribute, the text shown on hover.
width	number	Filled when the img tag states a width; many pages leave the size to CSS.	50
height	number	Filled when the img tag states a height; many pages leave the size to CSS.	50
format	text	Read from the file extension, for example jpg or webp; 'data' for an image embedded in the page.	svg
isLazy	true/false	Whether the page loads the image late, through loading=lazy or a data-src placeholder.	False
lazySource	text	The real image behind a lazy placeholder; empty when the image loads normally.
srcsetCount	number	How many alternative sizes the page offers for this image; 0 when it offers none.	0
srcsetLargest	text	The largest of those alternatives, which is usually the one worth downloading.	https://thumb.wikimedia.org/wikipedia/en/thumb/9/99/Quest...
inLink	true/false	Whether the image sits inside a link, which makes it a button rather than an illustration.	True
linkUrl	text	Where that link goes; empty when the image is not inside one.	https://en.wikipedia.org/wiki/Main_Page
domain	text	The domain the image is served from, which shows whether a site hosts its own images.	en.wikipedia.org
isInternal	true/false	Whether the image is hosted on the same registered domain as the page.	True
occurrences	number	How often the same image URL appears on the page; a repeated logo counts once with a higher number here.	1
pageTitle	text	The title of the page the image was found on.	Web scraping - Wikipedia
pageImageCount	number	How many images the page has in total, repeated on every row.	10
pageMissingAltCount	number	How many images on the whole page have no alt text, repeated on every row so one row carries its own context.	3
pageDataUriCount	number	How many images are embedded in the page itself rather than loaded; these are skipped unless includeDataUris is on.	0

Error codes: INVALID_URL, ROBOTS_DISALLOWED, NOT_HTML, NO_IMAGES, FETCH_FAILED, BLOCKED.

One delivered row looks like this:

{
  "pageUrl": "https://en.wikipedia.org/wiki/Web_scraping",
  "position": 1,
  "url": "https://en.wikipedia.org/static/images/icons/enwiki-25.svg",
  "altText": null,
  "hasAlt": false,
  "title": null,
  "width": 50,
  "height": 50,
  "format": "svg",
  "isLazy": false,
  "lazySource": null,
  "srcsetCount": 0,
  "srcsetLargest": null,
  "inLink": true,
  "linkUrl": "https://en.wikipedia.org/wiki/Main_Page",
  "domain": "en.wikipedia.org",
  "isInternal": true,
  "occurrences": 1,
  "pageTitle": "Web scraping - Wikipedia",
  "pageImageCount": 10,
  "pageMissingAltCount": 3,
  "pageDataUriCount": 0,
  "status": "ok"
}

Related actors from steadydata

• page-links-extractor: the links of the same page, with anchor text
• image-ocr-text: the text inside the images you found
• structured-data-extractor: the page's own structured data and Open Graph

Pricing

Pay per event: one image-listed event per delivered result. No charge for inputs that fail, no separate platform-usage surcharge.

Free Apify plan: this actor delivers up to 25 rows per run for accounts on the Apify free plan, and then stops with a message. That limit is set by us, not by Apify. It exists so the actor keeps paying for itself for the people who do pay. Any paid Apify plan runs it at full size, billed per delivered row, with failed rows never charged.

Reviews: if this actor saves you time, a short review on this page is the one thing that helps most. Ratings are what other buyers look at first, and we have no other way to ask.

FAQ

Is personal data collected? No. Image URLs and alt texts are what the page publishes; the actor reads only the URLs you give it and downloads no image.

How do I find images without alt text? Filter on hasAlt false. altText is empty both when the attribute is missing and when it is an empty string (decorative images); hasAlt is false in both cases, and pageMissingAltCount on the page row gives the total.

Which URL is delivered for a lazy-loaded image? The real source: when src is empty or a placeholder data: image, the lazy attribute (data-src, data-lazy-src, data-original and the like) or the first srcset candidate is used, and lazySource shows it. srcsetLargest gives the biggest candidate for high-resolution needs.

Can it download the images? No, by design; it lists them. Feed the url column to a downloader of your choice.

What does a run cost when a page fails? Nothing. ROBOTS_DISALLOWED, NOT_HTML, NO_IMAGES, FETCH_FAILED, INVALID_URL and BLOCKED rows are free; only delivered images are charged.

What happens when the source changes? Sources change from time to time; that is the nature of this work. The actor is monitored daily and fixed fast, and while it is broken you are not charged, because only delivered results cost anything.

Can I call it from my own code, like an API? Yes. Every run is available through the Apify API: start it with one HTTP request, then read the results as JSON, CSV or Excel from the dataset endpoint, or use the official Python and JavaScript clients. The input is the same JSON as in the input example above, so a scheduled script, a Make or Zapier step or an AI agent can call it without any changes.