Wind forecasts vs thermal forecasts: the distinction most people gloss over
XCWeather is great for wind. It is not a soaring forecast. The distinction matters more than most pilots realise, and most apps still get it wrong.
XCWeather has been running for over twenty years and is, by some distance, the most-used wind forecast site in UK paragliding and sailing. It is free, ad-supported, mobile-friendly, and reliable. Most people I fly with have it bookmarked. So the question of 'why would you build something to compete with XCWeather' has a short answer - we are not. The longer answer is that wind forecasts and soaring forecasts are different products solving different problems, and the distinction is worth pulling apart.
A wind forecast tells you what is going to happen at the surface. Speed, direction, temperature, pressure, visibility - the things that decide whether a hill is flyable at all. For a paraglider pilot driving to a site, this is most of what they need to know in the morning. Will the wind direction work for this launch? Is it under the upper limit? Is there a sea breeze coming in? XCWeather answers those questions cleanly, with familiar units and a UK-Europe map you can read at a glance.
A soaring forecast is a different question entirely. Given that the wind is flyable, how high will the lift get me, how strong will the thermals be, when will they start, when will they stop, and where will the cloudbase end up? Those answers need a vertical profile of the atmosphere, not just a surface measurement. They need a PBL scheme that has been thinking about how the boundary layer grows through the day. They need cloudbase calculated from the moisture profile, not assumed from observation. They need a thermal updraft velocity field that is derived from surface heat flux and stability, not extrapolated from wind.
XCWeather does not try to be a soaring forecast. It is a wind-and-weather observation/forecast site driven by public sector data (UK Met Office, under the Open Government Licence), with no developer API and no claim to be tracking thermals. That is fine. The problem is what happens when a developer building a paragliding app reaches for it as the data layer, hits 'no API', and falls back to scraping or to a general weather API that also is not a soaring forecast. The pilot using that app then thinks they are looking at thermal predictions when they are looking at extrapolated surface wind.
The honest split: use XCWeather (or its equivalents) to decide whether to drive to the hill. Use a soaring forecast to decide what to do once you are there. The two answers come from different models, with different underlying physics, and trying to make one do the other's job is where most pilot-app weather data goes wrong.
Convek is the soaring half. We run a WRF model four times a day on GFS data, downscale it to 4 km over specific countries, and serve the derived soaring fields (wstar, BL height, BL top, MSL cloudbase, convergence) as a JSON API. We do not try to be XCWeather - the wind-at-the-surface job is well covered, and our pipeline is optimised for a different question. The side-by-side comparison spells out where each one fits, and where neither is the right answer (turns out 'pilot UI showing soaring fields nicely' is not what either of us does - that is what SkySight, TopMeteo and the comparison index are for).
If you are a pilot, keep using XCWeather - it is a community staple for good reason. If you are a developer building a paragliding or gliding tool and you have been using a wind feed as the only weather data, that is the half of the picture you have been missing.