KENSAI Product Update: Locale-Split JSON and Overview Parity Tighten Same-Day Publishing

Today’s product work was about killing quiet drift. Every required language now needs its own JSON lane, its own localized article surface, and the same-day overview parity that stops English from sprinting ahead alone.

What shipped today

This update is less flashy than a new dashboard, and more important. We tightened the publishing path so multilingual blog output stays structurally honest under pressure instead of looking complete while the locale layer quietly drifts.

The core changes are simple: each language owns its own JSON index, same-day overview pages are expected to match English for the required locales, and HTML files remain the source of truth for rebuilding the derived indexes.

1) Each locale now owns its own JSON index

The hard rule is now explicit in the product surface: English posts belong only in blog-posts.json, German only in blog-posts-de.json, French only in blog-posts-fr.json, and so on. Mixing languages inside the English index is not multilingual publishing, it is metadata corruption with nicer branding.

That split matters because every downstream surface depends on it. Filters, overview cards, hreflang expectations, and same-day count checks all become less trustworthy the second one locale leaks into the wrong JSON file.

2) Same-day overview parity is no longer optional

We also tightened the expectation that the same-day blog overview for each required language must reflect the same publishing set as English. If English has the day’s product update, the required locale overviews should show that post on the same day too, not after a quiet lag or a manual cleanup pass.

That parity is not a cosmetic SEO detail. It is the difference between a truly localized publishing system and one that only pretends to be synchronized when someone checks the homepage quickly.

3) HTML stays the source of truth, derived indexes get rebuilt from it

Another important tightening is procedural: HTML article files remain the source of truth, while the JSON indexes are regenerated from those real pages. That gives the publishing path a cleaner recovery mode when JSON gets polluted, duplicated, or simply falls behind the actual shipped content.

It is a better contract. Real pages first, derived indexes second, overview generation after that. The order sounds boring, which is exactly why it works.

Why this matters

Same-day multilingual publishing breaks in boring ways: one language lands in the wrong index, one overview misses the latest article, one locale page exists without matching metadata, and suddenly the system looks complete only from one angle. This update is meant to close those gaps.

KENSAI should not publish like a demo. It should publish like a system that expects to be inspected.

What comes next

The next step is more of the same discipline: fewer silent drifts between HTML and JSON, clearer locale-level verification, and faster proof that all required language surfaces actually match on the day they ship.