Download the images from a marketing or landing page

Half the images on a major marketing page can be invisible to a naive scraper. On apple.com, 21 of 42 existed only inside noscript tags — the fallback a page offers readers without JavaScript — and reading those took the result from 50% to 100%. Paste a landing-page URL and the scanner reads all the places marketing pages actually put images: CSS backgrounds, retina variants, and those noscript fallbacks.

What this does

Measured on apple.com, where 21 of 42 images existed nowhere else in the served HTML — only inside noscript tags — and reading them took the result from 50% to 100% of what a real browser loads.
Hero and section art declared as a CSS background is read too, from inline styles, from style blocks, and from the stylesheets the page links — which is where marketing pages most often put the big image.
Retina and responsive copies of one image collapse into a single tile standing for the largest, so a page built at 2x does not read as twice the page.
Nothing is pre-filtered when you arrive here, on purpose. A hero can be a CSS background, a srcset photo or a noscript fallback, and every filter that would tidy the grid hides one of those three.

Where marketing pages put images

More places than almost any other kind of page. A single screen might carry a hero declared as a CSS background, a product shot as a srcset with four widths, an icon row as inline SVG, a social preview in a meta tag, and a plain copy of the hero inside a noscript block for readers without JavaScript.

That spread is the reason this page applies no filter when the grid opens. Every other tool page here narrows the view on arrival because the target is predictable. On a marketing page it is not.

What the noscript number means

The measurement behind this page is worth stating plainly, because it is the clearest evidence in the project that reading noscript matters. Against browser ground truth, apple.com scored 50% — half its images were simply not being found. Every one of the missing 21 was sitting in a noscript block that the parser was discarding.

Reading them took the same page to 100% with no other change. That is not a clever trick; noscript is content the site wrote for readers without JavaScript, and a static parser is precisely that reader.

What it won't do

Three linked stylesheets are fetched per scan. A background declared in a fourth is missed — and since the CSS claim above leans on stylesheets, this is the caveat that goes with it.
Marketing pages are the most animation-heavy class on the web. Scroll-triggered swaps, carousels and anything a script paints after load leave nothing in the HTML, so they are not found.
Brand and campaign artwork does not become licensed by being downloadable. A logo or a product render belongs to whoever made it, and a competitor's landing page is the least safe place to assume otherwise.
One site was measured, and apple.com is an unusually large and unusually carefully built marketing page. A smaller one may sit lower.
If the site's robots.txt asks automated tools not to read the page, the scan stops and says so. There is no override.

Questions

What is a noscript image, and why does it matter here?

A noscript tag holds what a page shows a reader whose browser runs no JavaScript. Marketing pages often put a plain copy of the hero imagery there while the scripted version handles everyone else. Most scrapers discard noscript because browsers do; a static reader is exactly the audience it was written for, so it is read through the same pipeline as the rest of the page.

The hero image is missing. Why?

Two likely reasons. It may be declared in a fourth or fifth linked stylesheet, past the three fetched per scan. Or it may be painted by JavaScript after load, in which case it is not in the HTML at all. A quick check is to view the page with JavaScript disabled — what remains is roughly what this can see.

Why is nothing filtered when I arrive from this page?

Because on a marketing page the image you came for could be in any bucket. Other pages here arrive pre-filtered, which is useful when the target is one format or one kind of tag. Here it would hide something, so the filters are left to you.

Can I get the 2x version?

You already have it. Each tile stands for the largest copy the page declares, so selecting it takes the retina version rather than the one your screen happened to load. The chip opens the smaller sizes if you want one of those instead.

Images belong to their creators. Only download what you have the right to use. See all tools or read what our requests do.