Release 1.14.5 · Richer docs libraries, safer rollout for embeddings, and clearer brand blog URLs

Documentation libraries gain optional per-page FAQs you can edit in the dashboard and readers see on published doc pages. FAQ text is included when we build search documents, so answers surface in org search alongside page body and description.

Library sections can opt into vector embeddings only when your organization is allowed: the product now treats docs.library_embeddings as an explicit org AI feature that is not implied by the usual experimental environment lists. Operators grant it per org from admin tools; dashboard UI explains stale embedding state so editors know when to refresh.

The brand blog area keeps a stable /blog entry point while the post list lives on a dedicated posts route, aligned with sidebar targets. Blog config follows the same experimental access rules as posts so the experience stays consistent. Org-level blog authors and categories pages are wired for the library navigation pattern you use elsewhere.

We tightened doc page updates so empty patches are rejected, optional metadata epochs can detect conflicts during concurrent edits, and new domain errors map to clearer API responses.