Loading page
About this tool
Review an image's intended purpose, alt state, nearby text, and longer alternative with calibrated W3C-aligned writing cues.
The Alt-Text Quality Assistant checks alt text against the thing that actually decides whether it is right — the image's purpose in this exact context — by asking you to classify it as decorative, informative, functional, complex, text-in-image or undecided, record the current alt state (missing, empty or present), and enter the essential information or the control's action. It then runs W3C-aligned rules over what you entered: empty alt on a meaningful image, a generic value like "image of", a filename used as alt, alt that duplicates nearby text, a complex image with no longer equivalent, and a word-overlap check against the information you said the image carries. It reviews text and context only — it never looks at the pixels, and a clean result is not a claim of WCAG conformance.
Open Alt-Text Quality Assistant on AltFTool — it loads instantly in your browser.
Choose the image's purpose in this context — decorative, informative, functional, complex, text-image or unclear — and the current alt state.
Paste the exact alt value, the essential information users need and any nearby visible text or caption, then press Review entered alt.
Read the outcome with its error and review-cue counts and each finding, copy the draft built from your own notes, or download the counts-only report.
The same photo needs different alt in different places, so the review is driven by the six purpose types rather than a generic style checklist.
Findings are graded error or review, and the outcome is one of needs-change, needs-human-review, or no configured cue triggered.
It reports plainly that word overlap misses synonyms, tone and accuracy, and that zero findings does not establish conformance.
When it is purely decorative — it adds no information and performs no function that is not already available in text. Use an explicitly empty alt="" rather than omitting the attribute, because a missing attribute leaves screen readers to fall back on the filename. The assistant flags a missing attribute as an error and an empty alt on a meaningful image as an error too.
WCAG sets no universal character limit. This tool flags alt over 200 characters for review, not as a failure — the question it asks is whether a concise alt plus a longer equivalent nearby would communicate more clearly than one very long string.
Usually not. Screen readers already announce the image role, so the prefix is redundant; the tool flags it for review rather than as an error, because occasionally the distinction matters — a painting versus a photograph of the same scene, for instance. An alt that is only "image" or "photo of" with nothing after it is flagged as an error.
No — and this one says so. It applies rules to the purpose, state and context you enter, and its word-overlap heuristic requires 50% of the meaningful words you listed to appear in the alt (25% for complex images) before it stays quiet. It cannot judge synonyms, accuracy, tone or whether a description makes sense to a real user, and it never inspects the image itself.