Describe with AISA in the block editor (Gutenberg)

What it is

“Describe with AISA” writes the alternative text of an image without leaving the block editor. Select an image block and the button is right there: in the block toolbar, and in the sidebar next to the Alternative text field — where the ALT is actually written, and where it is most often missing.

One click is one AI call with your own key, on the same endpoint the Media Library uses: same generation, same cost, same result. The description is written into the attachment and into the block attribute at once, so the page shows the new text immediately instead of the old one until the next reload.

Alternative text is what screen readers announce, and WCAG 1.1.1 requires it: without it the image is invisible to blind users, and search engines cannot read it either.

Where to find it

In the block editor (Gutenberg), on any selected image block: the toolbar button, and the AISA — image description panel in the sidebar. On WordPress 7.0 and later the panel sits in the first sidebar tab, right under the Alternative text field; on earlier versions it goes into the default group, a little further down but just as usable.

The same generation is available elsewhere: Describe with AISA Vision in the Media Library (Edit Media screen, details modal, and the grid's Select mode for groups), and the whole-library run from AISA → Bulk SEO → Images or from 1-Click SEO with “Images (alt-text)” ticked. Images placed with Elementor and the Classic editor are covered through the Media Library and the bulk runs.

And it does not have to be a click at all: Describe new images on upload, in Settings → API Keys → AISA Vision, describes every image a few seconds after you add it — from the block editor too — so the alt field is already filled when you get to it. It is off by default and has its own article.

Requirements & tier

Free, on every tier, for one image at a time. It needs the upload_files capability and a configured AI provider — Cloud credits or your own key. Bulk alt-text for the whole library is the Premium part; the single image in the editor stays free.

The button appears only on images that live in the Media Library: a picture pasted from an external URL has no attachment ID, so there is nothing for the server to describe. With Vision on, the AI actually looks at the image (Claude, GPT or Gemini); with Vision off, or on a provider without a vision model, it falls back to the context — filename, title, caption, parent page.

How to use

Open the post in the block editor and select the image block.
Click Describe with AISA, in the block toolbar or in the sidebar panel.
Wait a couple of seconds: the panel says Describing the image… while the call runs.
The alternative text appears in the field, already saved on the attachment and applied to the block. Read it and edit it if you want — it is a starting point, not a verdict.
Save the post as usual.

Only the ALT is touched from here. The image title is never rewritten in the editor: the choice between alt only, title only or both lives in the Media Library, where a hand-written title can be protected on purpose.

Questions people ask

Why is the button missing on this image? Because the image is not in the Media Library — typically it was inserted from an external URL. Upload it to the library and the button appears.
How long is the text? Around 125 characters, descriptive, with no “image of” filler.
What language does it write in? The one detected from the title and caption, then the page language, then the site locale. On a multilingual site each image keeps its own language.
Does it overwrite an ALT I already wrote? Yes, the button describes on demand and replaces the field, so use it when you want a new description. If you change your mind before saving, the editor's undo brings the old text back.
Does it also change the image title? No. From the block editor only the alternative text is written; the title is handled in the Media Library.
How much does one click cost? One AI call — a fraction of a cent with your own key, one credit on Cloud.

AI alt-text with Vision
AI image titles
Supported editors and page builders