Alt Text Length Checker – Descriptions That Survive the Field
Paste the description you are about to attach to an image and this alt text checker answers two mechanical questions: does it fit the field you are typing it into, and does it contain any of the problems that can be spotted from the characters alone. Everything runs in your browser — nothing you paste is uploaded, stored or sent anywhere, and no image is ever asked for.
What alt text is, and who reads it
Alternative text is the description attached to an image for anyone who is not seeing it. Screen reader users hear it read aloud; braille display users read it under their fingers; it is also what appears when an image fails to load or is blocked on a slow connection. For those readers the alt text is not a supplement to the picture — it is the picture. A description that reads IMG_4821.jpeg tells them a file was attached and nothing else.
Why the field limit matters
Most platforms cap their alt-text field, and a description that runs past the cap is not shortened politely — it is cut, usually mid-sentence, and the remainder is simply never announced. That is the only reason this tool cares about length. A field limit is a truncation point, not a standard of quality: being under it is not a pass mark, and there is deliberately no recommended range, no ideal length and no traffic-light band anywhere in the tool. The preview shows exactly where the cut falls, computed on grapheme-cluster boundaries so an emoji, a flag or an accented letter is never sliced in half.
Empty alt is sometimes the correct answer
A purely decorative image — a divider, a spacer, a flourish behind a heading — carries no information, and alt="" tells assistive technology to pass over it silently. Writing decorative divider line into that attribute adds noise to every single pass through the page. Mark a row decorative in the checker and an empty description is reported as correct, with the reason stated; supply a description on the same row and you get a note, because an image cannot be both decorative and worth describing.
Why “image of” is redundant
A screen reader announces the element before it reads your text, so image of a dog on a beach comes out as “graphic, image of a dog on a beach”. Starting with the subject removes the stutter. The exception is a screenshot: when the fact that something is a screen capture is part of what the reader needs to know, saying so is a legitimate choice, and the tool words that finding as a judgement call rather than a mistake.
What a checker cannot do
No tool that reads a string can tell you whether a description is accurate, complete or useful — the only questions that genuinely matter. It cannot see the image, it does not know what surrounds it, and it has no idea which detail in the picture is the one carrying the meaning. That is why there is no score, no grade and no percentage here: a number would imply a judgement this tool is not in a position to make. The findings are named, mechanical checks with a status and a sentence each — length against the field, empty alt, filenames, redundant openers, alt that repeats the caption, strings with no words in them, links and line breaks. Clearing all of them means only that nothing mechanical is wrong. Whether the description is any good is still a question for someone who can look at the image.