MUSE Cell Biotech Platform

Backend architecture and bespoke AI imagery for a regenerative medicine platform — serving investors, clinicians, and partners without losing credibility with any of them.

Start a Similar Project →

Regenerative Medicine & Exosome Therapeutics

One of the more demanding design contexts there is.

Biotech is one of the more demanding design contexts there is. The audience is split between investors who need to feel confidence in the organization's trajectory, medical professionals who will scrutinize anything that looks scientifically hand-wavy, and partners evaluating whether this is a serious operation. Get the balance wrong in either direction — too clinical and it's forgettable, too stylized and it loses credibility — and the site works against you.

This project came in through an unusual path. Another development team had already produced a Figma file with the front-end direction established. Our scope was the backend architecture and all of the imagery — which meant the visual decisions we did own had to integrate cleanly with work we didn't originate, while still producing something cohesive.

At a glance.

3

Audience Types

Investors, clinicians, partners

Flux AI

Bespoke Imagery

Zero stock photos

Lottie

Process Animations

Cellular & therapeutic mechanics

Global

Research Footprint

Interactive map visualization

The image strategy

No stock library exists for this. So we built the imagery ourselves.

Regenerative medicine and exosome therapeutics don't have a great stock photo library — what exists is either too generic or too literal in ways that create scientific accuracy concerns. We used the Flux generative AI model to produce bespoke abstract backgrounds and stylized cell illustrations that gave the site a genuinely “future-biotech” aesthetic without resorting to anything misleading.

The boundary we held was deliberate: no generative AI was used for anything that could be interpreted as actual scientific reference — lab equipment, clinical photography, microscopy. Those either came from the client's real assets or weren't included. In a medical context, that distinction matters. The AI-generated work stays in the abstract and atmospheric layer, not the evidentiary one.

Abstract and atmospheric — not evidentiary. That's the line, and it held throughout.

The build

Lottie animations and interactive data for concepts that resist static explanation.

The backend runs on Next.js 15 with the App Router, React 19, Tailwind CSS v4, Framer Motion, and Lottie for motion graphics. The Lottie integration was particularly useful for visualizing abstract concepts — things like cellular processes and therapeutic mechanisms that are difficult to explain in static form. Custom Lottie animations turn what would otherwise be dense text sections into something that actually communicates.

The site carries a fair amount of interactive data visualization: a React Simple Maps integration showing the company's international research footprint, a pipeline visualization of therapeutic development stages, and a technology deep-dive module explaining the MUSE cell expansion and exosome isolation process in a way that a non-specialist can follow without it feeling dumbed down.

The stack.

Next.js 15 + React 19

App Router, full site architecture

Tailwind CSS v4

Design system integration with existing Figma direction

Framer Motion

Scroll-triggered animations throughout

Lottie React

Motion graphics for cellular and therapeutic processes

React Simple Maps

Interactive international research footprint

Flux (Generative AI)

Bespoke abstract backgrounds and cell illustrations

What it demonstrates

Working within constraints you didn't set. Responsible AI decisions in a regulated context.

This project covers a few distinct things: working within someone else's design system rather than originating one, making responsible decisions about AI-generated imagery in a regulated industry context, and handling technically complex subject matter in a way that serves multiple audience types at once.

The clearance issues that ultimately took the site down were unrelated to the work — it was a regulatory path problem on the US market side, not a product one. The platform did what it needed to do while it was live.

Services used in this project