Posts quoted as sources, embedded at the end of the article
When a source of the brief is a post — X, Instagram, TikTok, YouTube, Facebook — it is now added at the end of the generated article, in the order of the sources, with its rich rendering. This is a layout decision, not a writing one: no AI is involved, the writer moves the block where they want or deletes it. A post already placed is never added twice, even after a regeneration; context articles and media are never embedded.
The caption carries the credit
The media library is down to two text fields: the alt text, which describes the image, and the caption, which situates and credits it — “Context of the photo. © Author / Agency”, like a newspaper caption and like WordPress. The separate “credit” field is gone; existing credits have been merged into captions.
What the credit did on its own, the caption now does: the attribution of a stock image is written as the caption on import, a collage’s caption is assembled from those of its photos, an adaptation inherits its own, an in-place replacement keeps it, and the library search reads it. On the WordPress side, the alt text feeds the alt and the title, the caption feeds the caption — nothing goes into the description any more.
The caption on the image block
In the editor, the library caption pre-fills the image block on insertion or replacement, then is edited in place, under the image, for that article only. When set, it is exported as a figure with its figcaption — and as a native WordPress image block, with its caption; when empty, the image goes out bare as before: nothing published changes unless it was written.
The Planner only follows taxonomies that say what an article is about
The agenda’s “To check” list sometimes showed eighty articles that shared a section. But a section says where an article is filed, not what it is about. Every taxonomy now has a “Tracked by the Planner” setting: hierarchical sections are not, flat tags are, same rule on WordPress import, fixable in one click. Consistency conflicts and the event term picker only look at tracked taxonomies.