# MeteoModels > Free weather forecast model comparison tool — compare GFS, ECMWF, ICON, AROME, ARPEGE, > HRRR, MeteoBlue, UKMO, GEM, HARMONIE and more side by side, in real time, for any location worldwide. URL: https://meteomodels.vercel.app/ About: https://meteomodels.vercel.app/about.html ## What is MeteoModels? MeteoModels is a browser-based weather application that shows predictions from up to 10+ numerical weather prediction (NWP) models simultaneously for any location worldwide. Instead of relying on a single forecast, users see all major models at once — including temperature, precipitation probability, wind speed, humidity, UV index and more. It is free to use, requires no account or sign-up, and works as a Progressive Web App (PWA) installable on any device. ## Weather Models Compared | Model | Organisation | Resolution | Coverage | |--------------|---------------------------|------------|------------------| | GFS | NOAA (USA) | 0.25° | Global | | ECMWF IFS | ECMWF (Europe) | 0.1° | Global | | ICON Global | DWD (Germany) | 0.1° | Global | | ICON EU | DWD (Germany) | 0.0625° | Europe | | HRRR | NOAA (USA) | 3 km | CONUS (USA) | | AROME HD | Météo-France | 1.3 km | France + borders | | AROME 2.5 km | Météo-France | 2.5 km | France + borders | | ARPEGE EU | Météo-France | 0.1° | Europe | | MeteoBlue | MeteoBlue (Switzerland) | — | Global | | UKMO | UK Met Office | 0.09° | Global | | GEM | Environment Canada | 0.15° | Global | | KNMI HARMONIE| KNMI (Netherlands) | 5.5 km | NW Europe | | DMI HARMONIE | DMI (Denmark) | 5.5 km | N Europe | Ensemble probability plumes are available for GFS ENS (30 members), ICON ENS (39), ECMWF ENS (51), GEM ENS (20) and AROME HD (deterministic reference). ## Key Features - **Multi-model ensemble**: up to 10+ NWP models compared side by side - **Location-adaptive model weighting**: models are automatically scored and weighted by accuracy for the requested location — high-resolution regional models (AROME for France, HRRR for USA, HARMONIE for N/W Europe) receive higher weights when they cover your area - **Real-time PWS observations**: actual measurements from the nearest Weather Underground personal weather station within 2 km — temperature, humidity, wind, pressure, UV - **Multi-station rain consensus**: cross-validates rain detection across all active weather stations within 2 km; reports `station_rain` (boolean), `rain_confidence` (0–1), `stations_checked` and `stations_raining` - **Official weather alerts**: MeteoAlarm (Europe) and NWS (US) alerts filtered to exact location - **Clothing advice**: practical daily outfit recommendations in 5 languages, based on ensemble - **Hourly and 14-day daily forecasts** with precipitation, wind, humidity, UV - **Ensemble probability charts**: visualise forecast uncertainty across 141+ model members - **Interactive map** (OpenStreetMap) and nearby webcam view - **PWA**: installable on iOS/Android, works offline after first load ## Public REST APIs ### Model Ensemble API ``` GET https://meteomodels.vercel.app/api/models?lat=LAT&lon=LON ``` Returns a weighted multi-model ensemble for the given coordinates. Selects the best models for the location, applies location-adaptive weights, and returns hourly + daily forecast data. Cache: 30 min. Rate limit: ~1,250 calls/day (Open-Meteo free tier). ### Real-Time Observation API ``` GET https://meteomodels.vercel.app/api/observation?lat=LAT&lon=LON ``` Returns current observations from all active Weather Underground PWS stations within 2 km, including a multi-station rain consensus. Fields: `station_rain`, `rain_confidence`, `stations_checked`, `stations_raining`, full `observation` block per station. Cache: 5 min. Rate limit: ~8,300 calls/day (WU free tier). Both APIs return JSON and support CORS. No authentication required. ## Languages Catalan · Spanish · English · French · German ## Data Sources - Open-Meteo (open-source weather forecast API, https://open-meteo.com) - Weather Underground / The Weather Company (PWS real-time observations) - MeteoAlarm (EU official weather warnings, https://www.meteoalarm.org) - NWS (US National Weather Service alerts) - OpenStreetMap / Nominatim (geocoding) - Open-Meteo elevation API ## Frequently Asked Questions **What is the difference between MeteoModels and a regular weather app?** Most apps show one forecast from one model. MeteoModels shows all major models simultaneously with a weighted ensemble average. You can see where models agree (high confidence) and where they diverge (uncertain forecast) — essential for planning around borderline weather. **How does MeteoModels weight models?** Each model gets a base score, a regional bonus (higher for high-resolution regional models covering your area), a resolution bonus (finer grid = better terrain capture), and a terrain sensitivity modifier (for mountainous locations). Final weights are normalised to 100%. **How accurate is the rain detection?** The `station_rain` flag in /api/observation is true if any PWS within 2 km reports precipRate > 0 mm/h in an observation less than 15 minutes old. The `rain_confidence` score (0.0–1.0) represents the fraction of all active nearby stations confirming rain. **Is MeteoModels free?** Yes — completely free, no account required. The project is supported by the community via Ko-fi (https://ko-fi.com/marcpeleja). ## Author Marc Peleja — https://meteomodels.vercel.app/about.html