How we decide whether today is flyable
Ask any weather app what the weather is over Hamburg and it will tell you: 18°C, cloudy, wind 17 km/h. Useful, and completely beside the point if what you actually want to know is whether you can put a drone in the air.
That is a different question. It has a different answer, and nobody was answering it — so we did.
Good, marginal, poor
Skydex reads four things against the limits that actually ground an aircraft, and calls the result good, marginal or poor:
| Condition | Marginal from | Poor from | Why |
|---|---|---|---|
| Gusts | 25 km/h | 40 km/h | A gust is what flips a light airframe. |
| Sustained wind | 15 km/h | 30 km/h | Eats range and battery long before it becomes unflyable. |
| Precipitation | 0.1 mm/h | 0.5 mm/h | Drizzle is survivable. Real rain is not, for most airframes. |
| Cold | — | −5°C | LiPo territory: voltage sags and the pack quits mid-flight. |
No black box. Those are the numbers, and they are the same numbers everywhere the verdict appears.
Why gusts outrank wind
A steady 25 km/h is an annoyance. You fly into it, you plan the return leg, you land with less battery than you wanted. A 25 km/h gust on a calm day is a different animal entirely: it arrives with no warning, from a direction the airframe is not trimmed for, and a light quad does not have the authority to argue with it.
So gusts are weighted above sustained wind in the verdict. This is the single place where our answer diverges most sharply from a general forecast, which usually leads with the average and buries the gust figure.
The cold limit is about the battery, not about you. At −5°C a LiPo's internal resistance climbs, voltage sags under load, and the pack can hit cutoff with what the telemetry still calls 40% remaining. Dress warmly and it is still a poor day to fly.
The same rules, at three scales
The interesting part is that one rule set answers three different questions.
Right now. Live conditions for the city you are flying over, in the HUD, updated as you go.
This week. A 7-day outlook on every city page, each day scored the same way.
This year. Run the same thresholds across daily climate normals and you get the share of days in a month that are genuinely flyable — which turns out to be a far more honest picture than average temperature. Hamburg reads 85% in March and 52% in January. London, Reykjavík and Wellington each tell their own story, and none of them match what a holiday-weather chart would suggest, because a holiday-weather chart is about sunbathing.
You can browse all of it from the flying weather page, or go straight to any city on the map.
Where the data comes from
Live conditions and the outlook come from a self-hosted Open-Meteo instance running the ECMWF IFS025 model. Ours, on our own server — not a shared public endpoint, which matters both for reliability and because the public one is for non-commercial use.
The month-by-month figures come from ERA5 reanalysis. Both are Copernicus data under CC BY 4.0.
The verdict on top of it is ours.
Also new: day and night
While we were in there — the sun now follows the real time where you are flying, if you want it to. There is a day/night toggle in the flight HUD, next to the conditions readout.
It is off by default, deliberately. Spawning over a city at 03:00 local into a black world, without having asked for it, reads as a broken game rather than as a feature. Switch it on and the sun rises and sets where it actually would.