Kuroma
Media Accessibility
What this check reads on your page
Images are listed first, leaving out tracking pixels and images marked decorative, then any video and audio elements. For images the check measures how many carry alt text, whether that alt is descriptive rather than generic, a filename or a couple of words, and how many information-bearing images sit in a figure with a caption; it also compares the number of images to the visible word count, marking down pages heavier than about one image per hundred words. For video it looks for a captions track, and for video and audio alike it accepts an on-page transcript or a transcript in VideoObject or AudioObject JSON-LD; it also notes images whose only source is a lazy-loading attribute. A page with no media at all is reported as unscored, since there is nothing to fix.
Our verdict on the evidence
Rescoped to the evidence No study shows alt text lifts AI citations, and major AI fetchers request no images. Kept at reduced weight for one honest purpose: a text alternative must exist for content that is locked inside media.
Vercel and MERJ AI crawler study (569M GPTBot requests: zero JavaScript execution, no image fetches)
The major AI fetchers request no images and run no vision model, so anything that exists only as pixels is missing from what they read. The factor carries a small weight because no study links alt text to citation; it is here so content locked inside media has a text form.
What it is worth
7% of On-Page Content, which is 1.8 points of the overall 100.
Not scored on data pages (rankings, statistics and paginated indexes): the factor is dropped from the weighted sum and from the denominator, so a page of that type is never charged for it.
How audited pages score on it
Based on 27,916 audited pages.
The median page scores 35, and the middle half of pages score between 26 and 44.
- Good (75 to 100)160 (1%)
- Moderate (50 to 74)3,542 (13%)
- Needs work (25 to 49)18,207 (65%)
- Critical (0 to 24)6,007 (22%)
87% of audited pages score below 50 on it.
Where it fails most and least
Industries that fail it most
- Passenger Transport / Transit: 95% of 124 audited pages fail it
- Food & Beverage / CPG: 92% of 1,650 audited pages fail it
- DTC / Consumer Brands: 92% of 938 audited pages fail it
- Supply Chain / Logistics: 92% of 336 audited pages fail it
- Manufacturing / Industrial: 90% of 2,780 audited pages fail it
Industries that fail it least
- AI / Machine Learning: 70% of 416 audited pages fail it
- InsurTech / Insurance: 75% of 118 audited pages fail it
- Wealth Management / Investment: 77% of 190 audited pages fail it
- Cybersecurity: 78% of 102 audited pages fail it
- SaaS / Cloud Software: 78% of 179 audited pages fail it
How to fix it
- Write alt text that states what the image shows and what it means; 'chart' or the filename scores as generic.
- Wrap charts and diagrams in a figure with a figcaption that gives the takeaway in one line.
- Publish a transcript beneath every video or audio embed, or add a captions track to the video element.
- Move any fact that exists only inside an image into the surrounding text.
- Give every image a real src attribute in the served HTML, not only a data-src that JavaScript fills in later.
What has changed in this factor
- Rubric version 2026-07-07: weight in its category moved from 15% to 7%