All products

Audit image alt text with AI

Review local HTML or image metadata for missing, generic or filename-like descriptions. Distinguish decorative images from informative ones before changing the page.

Audit image alt text with AI when a content team needs a manageable list of images to revisit. The accessibility-alt-text-auditor skill checks the supplied records for weak descriptions, suspicious lengths and missing alternatives. Its HTML helper can inspect image elements without retrieving the remote assets or running computer vision.

A product photograph labelled “image” probably needs attention, but its best alternative depends on the page. A purely decorative divider may correctly have an empty alternative. Use the findings to direct editorial review, especially when the same image source appears more than once in different contexts.

What it does

  • Locate missing alternatives in a local page.
  • Flag filenames and generic labels.
  • Check decorative images against the chosen policy.
  • Summarise coverage across supplied image records.
  • Review an optional proposed HTML copy before publication.

Install and use

  1. Install the skill and provide Python 3 plus an authorised local HTML file or image metadata JSON. A policy file can customise supported checks. No external account is needed.
  2. Ask, “Review these product images for missing or generic alt text. Separate decorative items and return a report first.”
  3. Check the actual image in its page context before approving a description. Use a separate output path if generating suggested HTML.

FAQ

Source

Made by sisodiabhumca. Original skill: accessibility-alt-text-auditor, published on GitHub under the MIT licence.

View on GitHub