Missing attribute
Find img elements that provide neither a text alternative nor an explicit decorative instruction.
Free image alt text checker / live HTML
Find missing alt attributes without flagging every decorative image.
Check one public page, inventory its image elements, and separate absent alt attributes from explicit empty values and non-empty descriptions.
✓ No signup✓ Live HTML✓ Read-only
50image details shown
2 MBbounded HTML response
3alt states separated
$0no account required
Attribute presence is objective. Image purpose still needs judgment.
Automated reports often merge missing and empty alt values into one inflated error count. RankLatch keeps them separate and shows the source, value, linked context, and relevant presentation hints.
Empty value
List alt="" separately so a person can decide whether each image is truly decorative.
Non-empty text
Expose the exact alt value without pretending that presence alone proves quality or accuracy.
Useful context
Show whether the image sits inside a link or declares a presentation or aria-hidden hint.
One accessibility signal—not an accessibility certification.
- ✓img elements present in the server-returned HTML
- ✓Missing, empty, and non-empty alt attribute counts
- ✓Exact bounded values plus linked and presentation context
- ✓Final HTTP response and redirect context
- —Whether the wording accurately communicates each image's purpose
- —CSS backgrounds or images added only after client-side rendering
- —Complete WCAG conformance or assistive-technology behavior
- —Automatic alt-text generation, ranking, traffic, or legal-compliance guarantees
Current inventory → change evidence
Know when a later edit adds images with no alt attribute.
RankLatch stores image-alt counts in each new approved page snapshot and records a notice if the missing-alt count later increases. It remains read-only and does not claim to judge the image or write the description.
See continuous monitoring →Straight answers
Before you check.
The right alt value follows the image's purpose in context. This tool identifies the states a public response can prove and leaves meaning to a person.
01What does this image alt checker inspect?+
RankLatch reads img elements in the bounded HTML returned by one public page. It counts missing alt attributes, empty alt values, and non-empty alt values, then shows the first 50 images in source order.
02Is an empty alt attribute always wrong?+
No. alt="" is the correct choice for many decorative images because it tells assistive technology to ignore them. An informative image generally needs useful alternative text. The checker separates empty from missing but cannot decide visual purpose for you.
03What is the difference between missing alt and empty alt?+
Missing means the img element has no alt attribute at all. Empty means the author explicitly supplied alt="". Those states can produce different accessibility behavior and should not be reported as the same thing.
04Does the checker grade the quality of alt text?+
No. It reports attribute presence and exact values. Good alt text depends on the image's purpose, surrounding copy, and destination when the image is linked; a character count or filename heuristic cannot prove that meaning.
05Does it execute JavaScript or inspect CSS background images?+
No. RankLatch checks the server-returned HTML. Images added only after client-side rendering and CSS background images are outside this bounded check.
06Does this replace an accessibility audit?+
No. A full accessibility audit includes keyboard behavior, semantics, labels, contrast, focus, forms, dynamic states, assistive-technology testing, and human judgment. This tool answers one narrower question.
07How does continuous monitoring use this signal?+
RankLatch stores the missing-alt count in each new approved page snapshot. After both snapshots contain that field, it records a notice when a later daily check has more images with no alt attribute. It does not generate alt text or judge image meaning.
08What happens on pages with many images?+
Totals cover every img element in the bounded response, while the detailed inventory shows the first 50. The page response itself is limited to 2 MB, and the result labels that boundary.