Live 3D bottle
A fully procedural three.js bottle (frosted-glass vessel, spun bronze cap, 150 instanced capsules visible through the frost) in one persistent canvas. Static-shot fallback when WebGL or motion isn't welcome.
Case study · Product · Brand site
An operator-built premium supplement brand. The waitlist site had to read as a fully-operating DTC (bottle-led motion, dose-disclosed editorial, a $49 founder tier capped at 500) before a single capsule shipped.
One week
Brief to live waitlist
6 + 3
Dose-disclosed ingredients on the panel, no proprietary blends
500
Founder-tier reservation cap, visible from the hero
0
Urgency timers, coupon codes, or fake testimonials
The product
IntelVit is a single-SKU nootropic brand for ambitious operators: a mushroom-led daily capsule packed in a frosted-glass cylinder with a bronze cap. The v1 site is the product: every section reads like the supplement-fact label, every scroll moment earns its place, and the waitlist is the only place a buyer can convert before the first batch ships.
Product in action
Screenshots from the IntelVit site: the July 2026 'The Site Is the Label' redesign, rendered in the brand's own Forest Studio system (cream paper, olive eyebrows, bronze doses) and embedded here so this reads as a real case study, not an abstraction.

Moment 1 · Hero
The frosted-glass bottle is a real-time three.js render, not a photo, and the milligrams are annotated right on the hero. No stock photo, no robot brain, no gradient mesh.

Moment 2 · Contents
Six chapters, six doses, each at the level the research actually used: name in serif, dose in bronze mono, mechanism in one sentence. This is the page label-readers screenshot and send to a friend.

Moment 3 · Terms
$49 a month for the first 500, locked for life. $79 a month after launch. The cap is real, the reservation bar is real, and there isn't a single urgency timer on the page. Trust scales when you stop using dark patterns.
The problem
Most nootropic brands market like 2014 wellness DTC: pastel gradients, urgency timers, influencer mosaics, and a marquee ingredient buried at a tenth of the dose the research actually used. Label-readers see through it in seconds. Operators do too.
IntelVit was founded by someone who actually takes the stack. The brief was unambiguous: the site has to read like the panel. Every milligram named. No proprietary blends. No countdown clocks. And it had to do all of that while still selling like a premium DTC brand to a buyer who has never bought a nootropic before.
Category sites lead with pastel gradients and influencer grids, none of which match a label-reader's frame.
Pre-launch DTC means no commerce backend day one. The waitlist has to be the product, without sounding like vaporware.
Most launches lean on urgency theater (timers, fake counts, limited-offer banners). IntelVit wanted the opposite: visible scarcity, named numbers, no dark patterns.
What we shipped
Each one ships in the current build. No flagged-off sections, no placeholders for things we never built.
A fully procedural three.js bottle (frosted-glass vessel, spun bronze cap, 150 instanced capsules visible through the frost) in one persistent canvas. Static-shot fallback when WebGL or motion isn't welcome.
Each of the six nootropics gets a dossier chapter: serif name, binomial, milligrams in bronze mono, mechanism. No 'proprietary blend' anywhere on the page.
Scroll unscrews the cap along its thread and sends capsules arcing out of the bottle: Lenis smooth scroll on the GSAP ticker, driving the same live 3D scene as the hero.
Peer-reviewed citations on a deep-olive full-bleed: study, dose, journal, year, for the exact doses in the capsule. Picked from peer review, not press releases.
$49 founder tier with an odometer count-up and a live reservation bar; the $79 launch price disclosed in the same breath. No coupon codes, no urgency timers, no 'request a quote.'
Email capture wired to Loops with idempotent handling, type-safe form action, soft-fail in dev, inline success and error states. No client-side state library.
How we shipped it
IntelVit was built the same way every IB Digital engagement runs: principal engineer driving, Claude Code translating intent to diffs, every section pair-reviewed against the brand brief. V1 went brief-to-live in a week on Next.js 16 + React 19, with Tailwind 4 CSS-first tokens so the Forest Studio palette lives in one place.
And it keeps moving at the same pace: in July 2026 the site was rebuilt ground-up as 'The Site Is the Label', a dossier-numbered one-pager wrapped around a live procedural 3D bottle, with Lenis + GSAP scroll choreography and react-three-fiber rendering. It shipped as a single release, in the same AI-native workflow as v1.
The result is heavy where it needs to be (a real-time 3D product, scrubbed scroll scenes, multi-chapter editorial) and quiet where it doesn't. The waitlist is a single server action; analytics is one client-side init; the email pipeline is one Loops API call. Production primitives, no toy libraries.
01
Every section composed in the same AI-native workflow we run for clients. Tight feedback loops on copy and layout, not weeks of slide reviews. Voice rules, banned-word lists, and palette tokens enforced at the file level.
02
GSAP + Lenis for the scrubbed open scene and masked line reveals; the live three.js bottle in one persistent canvas with adaptive quality tiers. A hard rule that nothing animates layout dimensions, and reduced motion always gets a settled page.
03
Server actions for the waitlist, Loops for email nurture, PostHog for conversion analytics, Vercel for deploys. The pre-launch site ships like a commerce site, ready to flip on Shopify the day the bottles arrive.
The outcome
Pre-launch DTC is a trap: most brands ship a placeholder splash page and call it a waitlist. IB Digital shipped the opposite. Every panel reads like the supplement-fact label, every scroll moment earns its place, and the founder-tier reservations started filling the same week the site went live.
Let’s talk
No slide deck, no discovery call fee, no ‘request a quote’ games. We’ll tell you whether it’s a fit in the first call — and point you somewhere useful if it isn’t.