SEO Tools·5 min read

Auditing Page Images: Alt Text and Bloat From One URL

Auditing page images sounds tedious because the manual version is: right click, inspect, scroll, repeat for fifty img tags. The Image Extractor from URL collapses all of that into one paste. Give it a URL and it returns every image on the page with its resolved address and alt text, plus counts of images with no alt attribute and images with an empty one. This article walks through a full audit built on that output: coverage, alt quality, file weight, and the mistakes that make image audits report problems that aren't really there.

Open the free Image Extractor from URL All insights

Why page image problems stay invisible

The page looks fine. That's the problem.

Images fail silently. A missing alt attribute changes nothing visually. A 3 MB hero photo still renders, just slowly, and on office wifi you'll never feel it. A product shot hotlinked from a supplier who might delete it tomorrow looks identical to one you host. None of this surfaces by looking at the page.

The only reliable move is an inventory: a flat list of every image address and its alt text, which is precisely what the extractor produces. Once problems are rows in a list, they stop being invisible.

Auditing page images: the one-pass inventory

Paste the URL, press Extract images, and read the badges before anything else. The first is the total, and it surprises people; pages you think of as having a dozen images often carry forty once icons, thumbnails, the og:image and inline CSS backgrounds are counted. The next badge counts img tags with no alt attribute, which is your accessibility to-do list in a single number. A separate badge counts img tags with an empty alt, which is usually a deliberate choice for decoration.

Filter chips split the list by where each image was found: img, lazy-loaded, picture/srcset, og:image, CSS background, inline SVG and icon. Tick Only images missing alt to narrow the grid to the real gaps. The grid previews the first 60 images, while Copy URLs and the Alt-text audit CSV cover all of them. The CSV holds each image's URL, where it was found, alt status, alt text, width and height, which is the file you want for the spreadsheet phase.

A worked image extraction on a product page

Input: one URL. Output: 41 rows.

Run a typical ecommerce collection page and the badges might read: 41 images, 8 img without alt, 4 img with empty alt. The eight are genuine misses: product photos whose cards show 'no alt attribute' where 'walnut standing desk, front view' should be. The four with alt="" are dividers and payment icons, and for pure decoration that is correct markup.

One row stands out for a different reason. Input: src='/cdn/hero-desk.jpg' resolves to Output: https://example-store.com/cdn/hero-desk.jpg, and opening that address reveals a 4200 pixel original serving a 600 pixel slot. That's the third audit category: weight nobody ordered.

Alt text audits without false positives

Alt text audits go wrong in two directions. The strict direction treats every empty alt as an error, and then someone 'fixes' decorative flourishes with text a screen reader must now announce. The loose direction writes alt='image' on everything and calls the badge satisfied. Both miss the point: alt text should say what a sighted user learns from the image, and stay empty when the honest answer is nothing.

The extractor keeps these cases apart. A missing attribute is marked 'no alt attribute', an empty one 'alt="" (decorative)', and images that aren't img tags, such as backgrounds and og:image, have no alt status at all. Work from the grid, where each image and its alt text sit side by side. That pairing is what makes the judgment calls fast.

Finding oversized images in the extracted URLs

Start with the copied URL list. Extensions tell the first story: .png files where .webp or .avif should be, or an /original/ path segment suggesting no resizing pipeline exists. Opening a few candidates directly shows their true dimensions, and anything several times larger than its display size is bandwidth spent on nothing.

One caution. For an image with a srcset, the extractor lists the largest candidate, so a big file there isn't automatically a problem; the browser may pick a smaller one for most screens. For file weight, paste a suspect URL into the HTTP Header Checker, where the content-length header usually gives its size in bytes.

Image audit mistakes that create false alarms

Four traps come up in almost every image review:

Counting decorative images with empty alt as failures, when empty alt is the correct markup for pure decoration.
Assuming every image lives in the HTML. The extractor finds inline style backgrounds and lazy-loaded data-src images, but backgrounds set in external stylesheets and images a script adds after load are not in the fetched source.
Writing alt text for og:image or CSS background entries, which aren't img elements and have no alt attribute to fill.
Stuffing keywords into alt attributes during the fix, which trades an accessibility gap for a spam signal.

Tips that speed up a page image review

Audit templates, not pages. One product page, one category page, one article: if the template has an alt gap, every page built on it shares that gap, and a single CMS fix clears hundreds of warnings at once. Extract again after the fix ships, because the missing alt badge dropping to zero is the cheapest regression test you'll ever run.

When you need the files themselves, for example to save your own images before a redesign or a host move, tick the ones you want and download them as a ZIP. If a preview won't render or a file won't download, don't assume the URL is wrong. Some hosts block hotlinking or browser downloads, so the address can be valid on the original page while refusing to load anywhere else. Those images are listed in a failed-downloads.txt file inside the ZIP.

Where the Image Extractor from URL hands off to other tools

The extractor inventories images; it doesn't judge the rest of the page. The Link Extractor does the same job for anchors when your audit moves from img tags to internal linking. The Broken Link Checker verifies that link targets actually resolve instead of returning 404s. And if the alt findings raise bigger questions, the WCAG Accessibility Checker tests the page against a full set of automated accessibility rules rather than one attribute.