About

Every TV station says its forecast is the one to trust. Weather Check writes down what each one forecast, waits, and checks: every forecast is scored against the official high, low and rain at the market's main airport, from the National Weather Service. The stations' weather teams and their websites are scored separately, alongside the National Weather Service and the main computer models. How it works, in detail: kcweathercheck.com/about.

Our reader

The site reads each station's public weather pages and forecast graphics, the way a viewer would:

• It identifies itself as weathercheck/1.0 (+https://weathercheck.us/about/) and sends from 66.85.7.19 (crawler.weathercheck.us).
• It reads each source once every 30 minutes and keeps only forecasts that changed.
• It reads a station's website only where the site's robots.txt allows it, and doesn't get around any site's blocking: a station that doesn't want to be read isn't.
• Also read: the National Weather Service's public data API, made for programs like this one (api.weather.gov); Open-Meteo's forecast API, made for programs like this one (api.open-meteo.com); the Storm Prediction Center's published outlook and report files (www.spc.noaa.gov); Gray stations' picture host: the forecast graphic a station's own weather page shows (webpubcontent.gray.tv); Hearst stations' picture host: the forecast graphic a station's own weather page shows (hips.hearstapps.net); the video host of FOX4 and other Nexstar stations: the forecast video a station's own article shows (lura.live). Their robots.txt files ask crawlers to stay out. They aren't crawled: only the published data, or the one graphic or video a station's own page shows, is fetched.

If you run a station's website and have a question, or want it to read less often or not at all, email hello@kcweathercheck.com.