Texada -- Store OS

TEST server. Nothing for a store runs here; this is the machine the till will talk to.

ledger database: reachable till sync endpoint: on

Support contacts

Open server status and this business's support contacts (office sign-in required). Keep a printed copy in the till-down kit for an outage.

Automated checks use GET /api/status.php?health=1: a read-only JSON health ping.

What this is

The server half of Texada's first proof: a till (the tablet at the counter) keeps selling when the internet drops, queues every sale locally, and replays the queue here -- POST /api/sync.php -- where each sale is checked against the laws (hours, the 30 g limit, the price floor, stock), written to the append-only ledger, and answered one by one. Test data only: a synthetic store, never a real sale.

What you can do here

• Read the five till screens as a first-day employee will see them: /till/ -- the words for the refund PIN, a cash mismatch, the 30 g stop, a crash restart, and the idle lock (M1 copy previews; the till's real look comes with M1).
• A browser GET of /api/sync.php answers 405 on purpose; only the till's POST with its token is accepted. Its read half is GET /api/status.php -- “did my last sale land?” -- same token, fails closed.

What's next

• The one-day tablet spike: ring a sale on the in-hand tablet, pull its power mid-write, watch it replay here.
• Before any real sale: the ledger moves to its own dedicated server (the production box), never this one.