AI image titles

What it is

Alongside the alt-text, AISA generates a short SEO title for each image (3–7 words, no quotes). When Vision is active it is based on what the AI sees; otherwise it is derived from the context (filename, caption, parent page). The title is only applied when the attachment still has its default title (the bare filename), so your hand-written titles are never overwritten.

Where to find it

Generated together with alt-text from the Images tab in Bulk SEO and from 1-Click SEO ("Images (alt-text)"). The image provider and Vision toggle live in Settings → API Keys → Images & Vision.

Requirements & tier

Same as Vision alt-text: a vision-capable provider (Claude / GPT / Gemini) for the "sees the image" path, or any text provider for the context fallback. Available in BYO-key and Cloud.

How to use

Configure the image provider in Settings → API Keys → Images & Vision.
Run the Images action in Bulk SEO (or 1-Click with images ticked).
AISA returns a JSON pair (alt + title) per image.
The title is written only if the attachment still has its automatic filename-based title.
Open the media item to review or edit the result.

Questions people ask

Will it overwrite my existing image titles? No — it skips any attachment whose title is no longer the bare filename.
How long is the title? 3–7 words, no quotation marks.
Does it need Vision? Better with Vision, but it still works from context if Vision is off.
Is the title in my language? Yes, it follows the same language detection as the alt-text.
Does it count as a separate operation? No — alt and title are produced in one AI call per image.

Troubleshooting

If a title is not written, the image probably already has a custom (non-filename) title — that is by design. If titles look unrelated to the picture, enable Vision and use Claude/GPT/Gemini so the AI can actually see the image.

AI alt-text with Vision · Product Image SEO Template · Product-images filter and the Images column.