Invisible until ready.

There's a rule the engine follows that we don't talk about much, because it's mostly the absence of a thing. A feature that depends on Vision data — a filter, a Smart Collection chip, an Editorial Layout hero pick — doesn't appear in the UI until the data it leans on is actually in the cache.

The internal name is ScanReadinessGate. It's a single coverage check that every Vision-dependent surface runs before it draws itself. If the gate is closed, the control isn't there. If the gate is open, the control appears, full and correct.

04. Why this matters

Most apps that lean on background analysis show every UI element from launch and then stub the ones that aren't ready yet — a filter that returns zero results, a button that says "no data," a tooltip explaining why you can't click. flexGrid hides them entirely.

The status bar shows an "Analyzing N/M" pill while the sweep is happening, so you can see the work. But the menus stay empty until the menus would actually mean something. The surface effect is that the app never lies about what it can do.

07. What we get for it

Calmer onboarding, fewer support questions, no half-empty filter lists pretending to be complete. It's the kind of choice you don't notice when it's working — which is, on reflection, the whole flexGrid thesis.