Minimal control page for speed measurement

This page carries none of the application shell. There is no shared Tailwind stylesheet, no React island hydration, no left sidebar, no Mediavine/Grow ad stack, and no analytics. It is static HTML with a handful of inline CSS rules and nothing else.

Why it exists

Our real content and tool pages render inside a height-locked app shell that loads one large Tailwind sheet as a render-blocking stylesheet, then hydrates several React islands (the sidebar, banners, the top bar) and, on ad-eligible visits, the third-party ad stack. Each of those layers costs something different: the stylesheet gates first paint, the ad scripts dominate the main thread, and the island JavaScript adds to total blocking time even though it loads after paint.

By stripping all of it away, this page establishes a baseline — the fastest a page can plausibly be on the same server and CDN. The gap between this baseline and a real page is the combined cost of the shell.

What we expect

With no render-blocking external CSS and no JavaScript at all, first contentful paint and largest contentful paint should land within a few hundred milliseconds even under simulated mobile throttling, total blocking time should be near zero, and cumulative layout shift should be exactly zero because nothing loads late to push content around.

If that holds, it confirms the infrastructure itself is fast and that the headroom on real pages lives entirely in the shared stylesheet, the islands, and the advertising scripts — not in the server, the database, or the network path.

There is nothing to interact with here. This route is excluded from search and linked from nowhere; it is purely a measurement control.