Open-source forecasting infrastructure, anywhere on Earth.
Our infrastructure powers high-resolution weather forecasting for energy operators and data platforms where other models stop–from remote villages to offshore wind farms.
High-resolution
Weatherwise operates open-source numerical weather models to deliver up to 72-hour forecasts at 1-km resolution, down to 250 m, with outputs every 15 minutes.
Domain experts
Many organisations already run WRF or other NWP models in-house, often at high cost and with limited focus. Weatherwise exists to take over that operational burden. We are a team of meteorologists and infrastructure engineers: We know the science behind the models we run, and the systems that make them efficient. That dual expertise is what lets us deliver transparent, operational forecasts at scale.
Built for complex regions
Most global or regional models stop at 2-9 km and cover only Europe or the US. Weatherwise runs sub-kilometer forecasts anywhere: urban area, desert, island, mountain, or offshore.
Coming nextTransparent infrastructure
Our mission is to build a shared infrastructure to run, compare and operate any open-source and proprietary numerical weather prediction models–physics and AI based.
Multi-model operation
Run any open-source NWP model (Meso-NH, WRF, ICON, UFS and more) anywhere on Earth with consistent configuration and output formats.
Unified model access
Fetch data from 20 + existing high-resolution models including AROME, ICON-EU or ECMWF IFS through a single API.
Model inter-comparison
Compare model outputs, compute basic skill metrics (bias, RMSE, MAE), and explore forecast uncertainty across models.
Efficiency transparency
We will publish runtime and energy usage trends for each configuration to drive down forecasting cost globally.
Coming soonUnified access
We partner with private and public institutions to give unified access to short, medium and long-range weather and climate prediction models.
1# pip install weatherwise
2from weatherwise import Client
3c = Client(api_key="YOUR_API_KEY")
4res = c.forecast.short_term(
5 lat=48.8566,
6 lon=2.3522,
7 variables=["temperature", "precipitation", "wind"],
8 end="2025-09-23T12:00Z",
9 models=["ecmwf-ifs-hres", "gfs"]
10)
11print(res)