In March of 2022 I was on a call with a race director two weeks before his event, trying to explain why the weather app on his phone wouldn't be enough to decide whether to start on time. He wasn't being careless — he'd used it for years, like every other trail organizer I'd talked to. But his route crossed three mountain ranges and climbed past 2,500 meters, and a single forecast icon for the whole region told him nothing about what his runners would actually face at the highest checkpoint, hours after the valley-floor start. That conversation is where MeteoTrail began, and it's the second product we've built at Maladeta Studio.
1. A forecast is not the same as ground truth
A consumer weather app, and most professional forecast products too, are built to answer a different question than the one a race director is asking. They tell you what the atmosphere over a region is likely to do — averaged, smoothed, and centered on a single reference point that might be a town in the valley. An ultra-trail route rarely stays in one place. It climbs, drops, crosses ridgelines, and spends hours at altitudes where temperature, wind, and precipitation behave nothing like they do 20 kilometers away and 2,000 meters lower. The gap between "forecast for the region" and "conditions at this exact checkpoint, right now" is not a rounding error. On a mountain course, it's the difference between a runner in a t-shirt and a runner in a whiteout.
2. What a race director actually needs to decide
Race directors aren't checking the weather out of curiosity. They're making decisions that can't be undone: whether to hold the start, whether to reroute around an exposed ridge, whether to evacuate a checkpoint before a storm arrives. Those calls depend on knowing, with enough lead time, what's happening at specific points on the course. A generic forecast can tell you it might rain somewhere in the mountains today. It can't tell you that lightning is already visible from checkpoint 4, or that the wind at the col has picked up past a safety threshold. That's the specific, checkpoint-level, near-real-time picture MeteoTrail set out to give organizers.
- Current conditions at each checkpoint: status, temperature, humidity and wind
- Automatic 48-hour forecast alongside a 7-day outlook
- Official alerts from national weather agencies, covering Spain, France, Portugal, Italy and more
- A single interface aggregating public webcams along the route
- Real-time lightning strike visualization
- A backoffice where a hired meteorologist can publish their own report directly to the whole organization
- Email notifications when conditions change
3. Building for events, not for everyone
MeteoTrail isn't a general-purpose weather product with a trail-running skin on it. It's a semi-automated weather-prediction tool built specifically for large outdoor events, and that shows in who it's for: ultra-trails, cyclotourism marches, Ironman races, ultra marathon MTB events, multi-stage tours, multi-day mountain activities, and even the regional safety monitoring for an entire comarca running several events at once. The pricing reflects that too — it's tiered by how long and how large the event is, with the option to hire a professional meteorologist through the platform itself when an organization needs a human making the final call, not just a dashboard. Building for a race director's actual decision, at the actual moment they have to make it, meant saying no to a lot of features a general audience would want.
4. What Val d'Aran by UTMB taught us early
One of our first real clients was Val d'Aran by UTMB, a demanding ultra-trail race through terrain where weather can turn in the time it takes a runner to cross a ridge. Their race director put it to me in a way I still think about: they can't control the weather, but MeteoTrail helps them make decisions based on it. Not a nice-to-have dashboard for the website — a tool that sits underneath the actual safety calls the organization makes during the event. That distinction is the whole reason MeteoTrail exists.
We've since built out a lot more of MeteoTrail's architecture — how it aggregates official alerts across four countries, how it decides when to escalate a notification, how the backoffice hands control to a human meteorologist without getting in their way. That's a separate post, for when I've had time to write it properly. For now, the honest open question is the one we started with: how much of a race director's safety decision can a tool like this actually take on, and how much should always stay with a person standing at that checkpoint, in that weather, right now?