An Image Checker That Reads the Page Like a Screen Reader

Accessibility overlays promise to fix alt text with a script. This tool does the opposite: it shows you every image the way assistive technology and crawlers encounter it, and what the file behind it really is, so you can fix the source.

What the Report Grades

Three sections, each marked pass, warning or fail, rolled into a score out of 100.

ImagesEvery img, image input, inline SVG and role="img" element gets a row with its accessible name and a verdict: missing alt, alt that is really a filename, generic alt like "image" or "logo", "image of" prefixes, alt over 125 characters, a decorative image that is the only content of a link, an SVG icon in an unnamed button, missing width and height, lazy loading on the hero, and in URL mode broken images, wrong content types and files over 300 KB.
Size and formatEach image file is fetched (the first 32 KB) to read its real format and pixel size; the weight comes from the response. Files over 300 KB, JPEGs and PNGs that should be WebP or AVIF, GIF animations, images served several times larger than their display box, an aspect ratio that disagrees with the width and height attributes, wrong content types, and the total weight of the page's images.
PageThe things around them: the lang attribute, the title, a main landmark, figures without captions and inline background images.
The rulesWCAG 2.2 (1.1.1 Non-text Content) for the alt text, Google's image guidance, and the usual web performance budgets for size and format.

The alt text guide shows what to write for each kind of image.

How a Page Becomes a Report

In URL mode, imagecheck.tools fetches the page once from this server, then fetches the first 32 KB of up to forty of the images it references to learn their status, content type, real format, pixel size and file size. The HTML and the probe results come back to your browser, where the analysis runs. In paste mode the HTML never leaves your browser and there are no probes. Both modes use the same analysis, which parses the document with the browser's own parser and walks every img, image input, inline SVG and role="img" element.

Each image gets an accessible name the way a screen reader would compute it (alt, then aria-label, then aria-labelledby, then an SVG title) and a verdict from a set of rules: missing, decorative, filename-as-alt, generic, prefixed, over-long, duplicated, a decorative image alone in a link, an unnamed SVG icon in a control. Technical attributes are read alongside: width and height, loading, fetchpriority, srcset, figure and figcaption. The file behind each image is identified from its first bytes, not its extension, so a PNG renamed .jpg and a JPEG served as image/png both show up. The rules come from WCAG 2.2 success criterion 1.1.1 and from Google's image documentation, and each finding says which rule it is applying.

What it cannot judge

Whether an alt text is a good description. "Woman smiling" passes every mechanical rule and may still be wrong for the page. The tool catches the cases that are wrong by construction (no alt, IMG_4032, "image", 300 characters) and leaves the writing to you; the alt text guide is the part that helps with that. It also does not run JavaScript, so images inserted by a script, lazy-loading libraries that write the src later, and single-page apps show as the HTML the server sent, not as the page you see. For those, save the rendered DOM from your browser's dev tools and use the paste tab.

When a verdict looks wrong

The contact page says what to send. For the rest of a page's indexability (title, canonical, meta robots), our sister tool indexcheck.tools picks up where this one stops.