Prime8

The situation

A hip-hop artist on the VisionApes Records label, and a brand where how the site feels is part of what is being sold. That is a real requirement, and it is also the requirement most likely to wreck a storefront, because motion is the easiest way to make a fast site slow.

The brief had four parts: a commerce storefront that does not feel like a theme, a motion layer that reads as designed rather than decorative, a gallery, and a link-in-bio page that does the job Instagram’s single link cannot. The releases, the videos and the storefront all sit on one domain, visionapesrec.com. The catalogue is up now and the shop opens when the merch drops.

Why headless here

We talk most merchants out of headless. It costs more to build, it costs more to maintain, and a well-built theme handles the overwhelming majority of storefronts better than the agencies selling headless will tell you.

The case here was specific: a content-and-commerce experience with a motion system that the theme layer genuinely could not deliver, on a brand where that experience is a real part of the purchase decision. That is a named reason. Without one, the answer is the theme.

Shopify keeps the checkout, the payments, the tax, and the orders. Only what the customer sees changes hands, which is what makes headless a manageable choice when it is warranted.

What we built

A Next.js 16 storefront built on the Storefront API, opening when the merch drops, with the checkout handed back to Shopify where it belongs.

A motion system on GSAP ScrollTrigger and Lenis: scroll-driven sequences, a custom cursor and interaction layer, and section transitions storyboarded before they were coded.

A contracted performance budget. The motion work had a byte and main-thread ceiling agreed before it was written. When a sequence exceeded it, the sequence changed, not the budget.

A gallery, image-heavy by definition, built around responsive sources and lazy loading from the first commit rather than optimized afterward.

A link-in-bio page on the brand’s own domain and design system, capturing traffic that would otherwise land on somebody else’s page.

Reduced-motion paths, built at the same time rather than bolted on at the end. Every sequence has a reduced-motion version that still reads as finished rather than merely unanimated.

What we deliberately did not do

We did not animate everything. One well-orchestrated sequence is worth more than twenty micro-interactions, and scattered motion is itself a sign that nobody made a decision.

We did not rebuild the checkout. Shopify’s checkout converts better than anything we would write, and it handles tax and compliance we do not want to own.

We did not use smooth scroll everywhere. Lenis is applied where it earns its place. Hijacking scroll on a product page is a good way to lose a sale.

Check it yourself

The site is live at visionapesrec.com. Open it on a phone, scroll through the motion sequences, then run any of its pages through PageSpeed Insights.

We are not publishing a conversion lift or a revenue number. We do not have clean attribution for one, and a number without a baseline, a timeframe, and a method is decoration. The site is live and it is measurable, and the storefront opens when the merch drops.