Real-time data-viz
Rendering live streams of data at sixty frames a second without dropping the rest of the page. Where the GPU helps and where plain canvas is the smarter call.
> // the lab
We experiment in the open.
The lab is where we push craft on our own time. Prototypes, motion studies, and AI experiments that have no client and no deadline, built to find what is worth bringing to real work.
> // the idea
What the lab is
A standing habit of building things just to learn how they behave, then folding what works into the studio.
01
Why we experiment
You cannot judge a technique from a blog post. We build the thing, break it, and find out where it is fast, where it janks, and whether it earns its weight.
02
No client, no deadline
Lab work answers to curiosity, not a brief. That freedom is what lets us take a real swing and find out what a technique can actually do.
03
Built to be measured
Every experiment gets the same scrutiny as client work: frame budgets, load cost, accessibility. A demo that only runs on our machine is not a result.
04
How it feeds client work
What survives the lab becomes a pattern we can reach for with confidence. The failures matter too, since they tell us what to never ship.
> // experiments
Recent experiments
A sense of what we explore. These are concepts we build to test, not live demos.
Generative motion systems
Motion driven by rules rather than hand-keyed timelines, so a system can feel alive and unique without a designer animating every frame.
Voice-agent latency tests
Measuring the real time from a person speaking to an agent responding, then hunting for every place to shave it. The gap between feels-instant and feels-broken is small.
On-device AI prototypes
Running smaller models in the browser to see what is possible without a round trip to a server. Faster, more private, and harder than it sounds.
Scroll-driven 3D
Tying a 3D scene to the scroll position so it stays smooth and degrades cleanly on weak hardware. The hard part is never the effect, it is the fallback.
> // the payoff
How it feeds the work
The lab is not a hobby on the side. It is how the studio stays sharp.
01
Sharper craft
Working at the edge of a tool teaches us what it does well. That judgment shows up in every client build, even the ones with no flashy effect.
02
Real speed
We learn the cost of a technique before it touches a client, so we can move fast without guessing. The hard lessons are already paid for.
03
Proven patterns
When we propose something ambitious, it is because we have already built it and watched it run. You get the result, not the experiment.
04
Honest limits
The lab tells us what does not work as well as what does. That is why we can say no to a trend with confidence and yes to the right thing.
> // lab faq
About the lab.
Sometimes. If an experiment has proven itself and fits your goals and performance budget, we will adapt it. We will not bolt a flashy effect onto a project just because it is ours.
The lab section describes concepts we build internally, not a gallery of public demos. When we do publish a working piece, we will say so plainly and link to it.
No. The opposite. Because we have already tested techniques in the lab, client builds move faster and carry less risk. The experiments are the prep work.
Because how we work is part of why the work is good. Showing the experiments is more honest than a list of buzzwords, and it tells you the craft is real.
> // operating principles
The non-negotiables.
01
Proof over promises
We show our work, our scores, and our reasoning. The site itself is the argument.
02
Restraint is the flex
One accent, one action, nothing competing. The discipline is what makes it premium.
03
Speed is a feature
Performance is not a tradeoff against beauty. We refuse to ship a site that janks.
04
You own everything
Your design, your code, your content. The retainer keeps it sharp, never hostage.
Want this kind of craft on your project?
The techniques we prove in the lab show up in client work, tuned to your goals and your performance budget. Tell us what you are building.