What's new?

A handful of dashboard fixes after the 2.0 launch — same look and feel, just better behaved at the edges.

• Compact view now shows 24 items per page (up from 20) so a full page actually fills the screen. Comfortable stays at 20.
• On very wide monitors the grid no longer fans out to 8+ columns of tiny cards, and on narrow screens it no longer collapses to a single column.
• The floating mascot has been retired from the dashboard.

Barcode label printing is now available to every workspace — no toggle, no waiting list. Existing workspaces are switched on automatically; new workspaces get it from day one.

2.0 also marks the official rollout of the new Stocklyst design language. The pieces have been landing for months — refreshed marketing site, live status page, faster login, redesigned reports — and 2.0 is the version that ties them together.

• Print labels for any item, any size, any time — straight from the inventory list.
• Pick from the built-in templates or design your own in the label builder.
• Bulk-assign a label layout across thousands of items in one pass.

Every marketing route now renders through a disciplined atoms → molecules → organisms → templates hierarchy. 35 reusable pieces (17 atoms, 11 molecules, 7 organisms), 6 route templates (legal, glossary, doc-article, compare, use-case, industry), one design-token layer. The page layouts you see — hero, compare tables, industry blueprints, pricing, FAQ — all compose from the same handful of primitives now.

• 34 routes migrated end-to-end (52 unique URLs including dynamic slugs).
• Every heading resolves through CSS variables — one token flip and the whole site follows.
• CI grep-check in GitHub Actions blocks future hex-color regressions inside marketing routes.
• All legacy page-shell files deleted — no more duplicated scaffolding across 30+ routes.

Stocklyst now ships a public status page with a Trunk-style 90-day uptime grid per service. The footer dot that used to claim "All systems operational" statically is now wired to live data — it turns amber or rose the moment a regional database probe degrades.

A cron runs every minute, probes each regional Postgres (EU / US / AP) with SELECT 1, classifies by latency, and writes the result to a central table. Public APIs serve a summary + 90-day history for the page and the footer.

Heading sizes across the marketing site were bumped to give real editorial weight. Hero now clamps to 7rem at desktop (was 6rem), section H2s to 3.5rem (was 2.75rem). DisplayHeading defaults to extrabold so hero panels carry the visual weight without every consumer opting in.

A subtle but consequential bug fix landed alongside: tailwind-merge was treating the custom text-m-* font-size utilities as conflicting with text-fg color utilities, silently stripping the color from every heading. Every heading on every page now renders the intended color in both light and dark sections.

A raw-SQL ON CONFLICT statement in the transfers route was silently dropping ~9% of destination inserts in the EU region. Rewrote the path to SELECT → UPDATE / INSERT via the Drizzle builder, then backfilled 108 affected rows. Multi-region transfers that previously looked "stuck at sent" should now close cleanly.