Five forecast engines vote on your weather. Every vote is graded against what actually happened. Atlas365 learns which engines to trust over your sky — and shows you the receipts.
Atlas365 consults the world's national forecast systems and the new AI models — then grades every one of them, per region, against observed reality. Trust is earned in the ledger, never assumed.
| Engine | Operator | Role |
|---|---|---|
| ECMWF | Europe — the global gold standard | VOTES |
| GFS | NOAA, United States | VOTES |
| ICON | DWD, Germany | VOTES |
| JMA | Japan — Asia & monsoon specialist | VOTES · ASIA |
| CMA | China Meteorological Administration | VOTES · ASIA |
| MF / UKMO | France & UK — European specialists | VOTES · EUROPE |
| GEM / BOM | Canada & Australia specialists | VOTES · REGIONAL |
| MET.NO | Norwegian Meteorological Institute | AUDITIONING |
| AIFS | ECMWF's neural-network model | AUDITIONING · AI |
| GraphCast | Google DeepMind, run by NOAA | AUDITIONING · AI |
| GEFS ×31 | 31 perturbed physics runs | UNCERTAINTY |
This sequence is the product. It runs on your phone for your places, and on our servers around the clock.
Up to five regional engines plus the AI models predict tomorrow for your exact coordinates.
When tomorrow arrives, every archived forecast is scored against what the sky actually did.
Engines that verify well over your region earn weight. Systematic bias is measured and corrected.
The forecast you see is Atlas365's own blend — no single engine's number, and nobody's copy.
Live multi-engine forecasts, Earth Watch hazards, area air quality and flood insight, AI briefings — and a scorecard that keeps us honest. Android build available now; Play Store listing on the way.