Hairline Stats Band
A four-up metrics strip with 1px hairline dividers, built from a single gap-px grid so the rules stay crisp on any display.
Performance figures with the measurement method printed beside each one, so a sceptical reader can check rather than take them.
components/stats-benchmark-band.tsxnpx hoverlab add stats-benchmark-band
Or over MCP, from your editor's agent — no account needed.
Free to read, copy and install, for personal and non-commercial projects. Shipping it in client work or a paid product needs Pro ($79 once). The source lands in your repo and stops being ours — no attribution, nothing to upgrade.
Start a page with this section — add more, order them, and leave with the page source.
Measured, not claimed
Every number below names the conditions it was taken under. A benchmark without its method is a marketing figure wearing a lab coat.
p50 over 10,000 invocations, Frankfurt, empty cache.
Single node, 4 vCPU, read-only workload.
One million rows, measured wall-clock end to end.
Peak RSS under the throughput run above.
Rendered live in your current theme — this is the same component whose source is below, not a screenshot of it.
/**
* <StatsBenchmarkBand> — Performance figures with the measurement method printed beside each one, so a sceptical reader can check rather than take them.
*
* A stat band's layout problem is that the figure and its caveat compete
* for the same slot. The obvious wrong answer is to drop the caveat: four
* huge numbers read beautifully and mean nothing, because "112 ms" without
* "p50, Frankfurt, empty cache" is not a measurement, it is a claim. So the
* detail line is part of the tile rather than a footnote under the section,
* and the tile is sized to hold two lines of it.
*
* The markup is a <dl>, and that is the non-obvious accessibility decision.
* A grid of divs reads to a screen reader as four unrelated numbers in a
* row; <dt>/<dd> pairs read as "Cold start, 112 ms", which is the only
* useful way to hear this section. The detail sits in a <p> inside the same
* <div> rather than in a second <dd>, because a definition list with two
* values per term announces them as alternatives.
*
* `tabular-nums` on the value, so the four figures line up on their digits
* rather than on their glyph widths — the column stops looking ragged the
* moment one value has a 1 in it.
*
* The grid is `gap-px` over a `bg-border` parent rather than four bordered
* cards. That produces true hairlines with no doubled edges at the seams,
* and it collapses correctly at every breakpoint without a per-cell border
* rule that has to know which column it is in.
*
* Server component. There is no state here at all, and the demo's numbers
* are the defaults so the block renders identically in a preview, a
* screenshot and someone's page before they replace them.
*/
export interface StatsBenchmarkBandMetric {
label: string
value: string
/** One clause on what moved, or omit for a bare figure. */
detail?: string
}
export interface StatsBenchmarkBandProps {
eyebrow?: string
heading?: string
intro?: string
metrics?: StatsBenchmarkBandMetric[]
className?: string
}
const METRICS: StatsBenchmarkBandMetric[] = [
{ label: "Cold start", value: "112 ms", detail: "p50 over 10,000 invocations, Frankfurt, empty cache." },
{ label: "Query throughput", value: "24.6k/s", detail: "Single node, 4 vCPU, read-only workload." },
{ label: "Index rebuild", value: "38 s", detail: "One million rows, measured wall-clock end to end." },
{ label: "Memory ceiling", value: "512 MB", detail: "Peak RSS under the throughput run above." },
]
/**
* A per-instance heading id, derived rather than generated.
*
* `React.useId()` is the textbook fix for a literal id in a reusable
* component and is not available here: this block has no `'use client'`
* and does not warrant one. It renders text, and a band that hydrates on
* every page to render text is a cost with nothing on the other side of
* it. Same call, and the same reasoning, as `footer-mega`.
*
* Hashing the heading *and* the intro is deliberate. Either alone works
* until two callers happen to agree, which is exactly what a catalog hub
* rendering every page into one document arranges for; between them they
* are the two props nobody leaves at the default.
*/
function headingId(heading: string, intro: string): string {
let hash = 0
const source = `${heading}|${intro}`
for (let i = 0; i < source.length; i++) {
hash = (Math.imul(hash, 31) + source.charCodeAt(i)) | 0
}
return `stats-benchmark-band-heading-${(hash >>> 0).toString(36).slice(0, 4)}`
}
export function StatsBenchmarkBand({
eyebrow = "Measured, not claimed",
heading = "How fast it is, and how we know",
intro = "Every number below names the conditions it was taken under. A benchmark without its method is a marketing figure wearing a lab coat.",
metrics = METRICS,
className,
}: StatsBenchmarkBandProps) {
return (
<section
aria-labelledby={headingId(heading, intro)}
className={`w-full bg-background px-6 py-16 sm:py-20 ${className ?? ''}`}
>
<div className="mx-auto max-w-5xl">
<p className="text-sm font-medium text-primary">{eyebrow}</p>
<h2
id={headingId(heading, intro)}
className="mt-2 text-3xl font-semibold tracking-tight text-foreground sm:text-4xl"
>
{heading}
</h2>
<p className="mt-3 max-w-2xl text-base text-muted-foreground">{intro}</p>
{/*
A <dl>, not a grid of divs. Each tile is a term and its value, and
a screen reader reading "Uptime, 99.98%" is the whole point of the
section — a div soup reads as five unrelated numbers.
*/}
<dl className="mt-10 grid grid-cols-1 gap-px overflow-hidden rounded-xl border border-border bg-border sm:grid-cols-2 lg:grid-cols-4">
{metrics.map((metric) => (
<div key={metric.label} className="min-w-0 bg-card p-6">
<dt className="break-words text-sm font-medium text-muted-foreground">{metric.label}</dt>
<dd className="mt-2 break-words text-3xl font-semibold tabular-nums tracking-tight text-foreground">
{metric.value}
</dd>
{metric.detail ? (
<p className="mt-1 break-words text-sm text-muted-foreground">{metric.detail}</p>
) : null}
</div>
))}
</dl>
</div>
</section>
)
}
bg-card, text-muted-foreground) — it inherits your theme instead of overriding it.Drop it at components/stats-benchmark-band.tsx and import it where you need the section:
import { StatsBenchmarkBand } from '@/components/stats-benchmark-band'2 of this block’s props are simple enough to drive from here. Change them and the block below re-renders — it is the same component whose source is above, not a mock of it. Everything else it accepts is in the table underneath.
Read out of the component’s own type and signature, so this cannot drift from the source below. Every prop has a default — the component renders standalone before you pass it anything.
| Prop | Type | Default |
|---|---|---|
eyebrow | string | "Measured, not claimed" |
heading | string | "How fast it is, and how we know" |
intro | string | — |
metrics | StatsBenchmarkBandMetric[] | METRICS |
className | string | — |
The same block rendered once to markup, wrapped as a file your framework compiles. Tailwind classes are framework-agnostic, so the design transfers intact — the behaviour does not.
This block is interactive. The markup below is its initial state with the event handlers stripped — you will need to re-wire the behaviour in your framework.
<!--
Benchmark Stat Band — markup from the Hoverlab catalog.
This is the block rendered once to HTML and wrapped as a component
file. It is not a port of the React source: the Tailwind classes carry
the design, which is the part that took the work, and they are the same
in every framework.
This block is interactive in React and the handlers are NOT here.
Buttons, toggles and menus render in their initial state and do
nothing until you wire them up.
-->
<section aria-labelledby="stats-benchmark-band-heading-220d" class="w-full bg-background px-6 py-16 sm:py-20 ">
<div class="mx-auto max-w-5xl">
<p class="text-sm font-medium text-primary">Measured, not claimed</p>
<h2 id="stats-benchmark-band-heading-220d" class="mt-2 text-3xl font-semibold tracking-tight text-foreground sm:text-4xl">How fast it is, and how we know</h2>
<p class="mt-3 max-w-2xl text-base text-muted-foreground">Every number below names the conditions it was taken under. A benchmark without its method is a marketing figure wearing a lab coat.</p>
<dl class="mt-10 grid grid-cols-1 gap-px overflow-hidden rounded-xl border border-border bg-border sm:grid-cols-2 lg:grid-cols-4">
<div class="min-w-0 bg-card p-6">
<dt class="break-words text-sm font-medium text-muted-foreground">Cold start</dt>
<dd class="mt-2 break-words text-3xl font-semibold tabular-nums tracking-tight text-foreground">112 ms</dd>
<p class="mt-1 break-words text-sm text-muted-foreground">p50 over 10,000 invocations, Frankfurt, empty cache.</p>
</div>
<div class="min-w-0 bg-card p-6">
<dt class="break-words text-sm font-medium text-muted-foreground">Query throughput</dt>
<dd class="mt-2 break-words text-3xl font-semibold tabular-nums tracking-tight text-foreground">24.6k/s</dd>
<p class="mt-1 break-words text-sm text-muted-foreground">Single node, 4 vCPU, read-only workload.</p>
</div>
<div class="min-w-0 bg-card p-6">
<dt class="break-words text-sm font-medium text-muted-foreground">Index rebuild</dt>
<dd class="mt-2 break-words text-3xl font-semibold tabular-nums tracking-tight text-foreground">38 s</dd>
<p class="mt-1 break-words text-sm text-muted-foreground">One million rows, measured wall-clock end to end.</p>
</div>
<div class="min-w-0 bg-card p-6">
<dt class="break-words text-sm font-medium text-muted-foreground">Memory ceiling</dt>
<dd class="mt-2 break-words text-3xl font-semibold tabular-nums tracking-tight text-foreground">512 MB</dd>
<p class="mt-1 break-words text-sm text-muted-foreground">Peak RSS under the throughput run above.</p>
</div>
</dl>
</div>
</section>
What each framework gets across the whole catalog — effects convert properly; this rung is markup.
The component, its props, the design tokens it expects and the command that installs it — as one prompt. Paste it into Claude, Cursor, v0 or ChatGPT and what they build around it will match the rest of the catalog instead of inventing its own system.
Want the whole screen instead of this one section? Open a page and copy it entire.
All of them free to read, copy and install — no account, no locked tiles, no watermarked preview. The whole catalog is open, and so are the API and the CLI.
Browse StatsCopying the code is free. Putting it in client work or a paid product is what Pro is for — the licence, not the access.
A four-up metrics strip with 1px hairline dividers, built from a single gap-px grid so the rules stay crisp on any display.
Metrics carrying the direction they moved, with the good direction declared per stat so a falling churn rate reads as a win — arrow and text, never colour alone.
Four figures in a hairline card next to the paragraph making the claim, each carrying an optional source line — for numbers that mean nothing without knowing lower than what, over how long, across how many.
The buyer’s current situation in one column and yours in the next, as a real table with scoped headers — the comparison they are actually making, which a delta against your own past cannot express.
The same measurement taken year after year down a vertical rail, each with the event that explains it — the shape for trajectory, which no snapshot of four numbers can claim.
The compliance and security facts a procurement reviewer asks for, stated as figures with their scope rather than as badges.