Pyrene operates its own metabolic testing lab in Toulouse. Athletes come in for a cardiopulmonary exercise test, run on a protocol designed by the team, and leave with a full physiological profile. The raw data produced by the lab machine is dense, clinical and not meant for athlete consumption. That is where the Metabolic Scan dashboard fits in.
Before the dashboard, interpreting a test meant opening a spreadsheet, cross-referencing reference ranges by hand, and sending the athlete a static PDF or a verbal briefing. The data was accurate but generic. Reading it required a specialist. And the athlete rarely kept it. Pyrene needed an output that matched the quality of the test itself.
One admin interface where the Pyrene team uploads each test file. One athlete-facing dashboard delivered through a private, secure link. Six sections of analysis. A dedicated design system. All of it built, deployed and running in one week from brief to beta.
No generative model. No black box. The interpretation layer is built on established sports science methods and custom algorithms that produce reproducible, auditable outputs. The athlete reads a report that a physiologist could validate line by line, delivered in the visual language of the Pyrene brand.
A testing service that deserved a proper output.
Pyrene had already invested in a full metabolic testing lab: machines, protocols, expertise. The bottleneck was at the end of the chain. Every test produced a rich dataset but the athlete only ever saw a fraction of it, filtered through a spreadsheet or a spoken summary.
Our brief was narrow and clear: build the missing piece. A product layer that turns a one-hour test into a lasting personal reference, delivered the day of the session.
The pipeline is deliberately simple on the outside. The Pyrene team uploads the test file through an internal admin interface. A custom parser reads the metabolic dataset, extracts the relevant markers, and structures them into a clean model. A private, secure link is generated and sent to the athlete. The report opens in a browser, on any device, at the quality level of a premium software product.
What is not on the outside is where the work lives. The algorithms that turn a raw oxygen consumption curve into training zones, fuel strategies and race pace estimates are the result of careful physiology, not of a language model making guesses. Every number in the report can be traced back to the data it came from.
Every report opens with the athlete's name and a one-sentence summary of the session. From there, the dashboard is organised as six chapters that read like a short book: overview, training zones, physiology, training plan, nutrition strategy, body composition.
Each section reframes the raw numbers as a concrete recommendation. Not "your VO2max is 71.0". But "you can hold 200 watts for an hour, here is what that means for your next race, and here is how to fuel it." The data stays accurate. The language finally matches the athlete.
The six sections follow a deliberate narrative arc. Overview sets the frame and the key number. Training zones translate physiology into effort. Physiology gives the full picture for the athletes and coaches who want it. Training plan turns the zones into a week. Nutrition converts effort duration into carbohydrate strategy. Body composition closes with the context that shapes everything else.
The dashboard lives under the Pyrene umbrella, so the design language had to extend the existing identity rather than invent a new one. Same typography, same restrained palette, same insistence on white space. A single accent colour signals interaction and draws the eye to the numbers that matter. Everything else stays quiet.
The same design system scales from desktop, where coaches and physiologists go into the details, to mobile, where athletes check their zones between sessions. One product, two reading contexts, one visual voice.
No AI. No black box.
Real physiology, reproducible by design.
The interface is full of small decisions that an athlete never notices if we did them right. A classification badge that only shows up when the data warrants it. A light mode toggle for athletes who read reports in bright sunlight. A sticky summary in the sidebar so that the key number stays in view while scrolling. Performance metrics that round to the precision the test actually supports, never more.
None of it is flashy. All of it protects the seriousness of the product.
Live in beta.
Shipping to real athletes.
One week from brief to beta.
The Pyrene Metabolic Scan is the kind of project we care about most: a clear problem, a short timeline, a product that replaces a spreadsheet with something athletes actually want to open. It extends a brand we already knew intimately, having built the Pyrene Performance identity system from the ground up, into a new territory. Same rigour, different output. Same language, translated from a shelf to a screen.
Florian B.
Co-founder — Pyrene Performance
"Before the dashboard, interpreting a test took us close to half an hour and still left the athlete with a static PDF. Today they receive their report on the day of the session and spend that half hour training instead. It changed the experience we can offer."
Emmanuel L.
Co-founder — Pyrene Performance
"Richard took a week to deliver what we had been putting off for months. The result looks like something an athlete wants to come back to, not a clinical export. That alone raises the bar of the whole service."
Pyrene®
French nutrition for ultra-endurance.
Forvis Mazars®
International audit, tax and advisory network.
Le Courti®
Guest house by the sea.
Menao®
B2B app for artisans and craftsmen.
Ultra Bike France®
French ultra-cycling event series.
Pace®
Triathlon and individual sports coaching.
Tell me about it — I'll get back to you within 24 hours with a free intro call proposal.
Not ready yet?
Browse more projects ↑