A convolutional ensemble resolves radar, satellite, and surface fields every 60 seconds — resolving storm cells before conventional NWP models finish one integration step.
Hyperlocal grid
Every forecast is native to a 0.25 km grid — tuned to terrain, canopy, and the microclimate at your exact coordinates rather than the nearest airport station.
Severe-weather lead time
Convective nowcasts deliver an average 45-minute lead on lightning, hail, and damaging wind — pushed to devices, dispatch, and incident command in a single hop.
Forecast Engine
Nowcasting, at the speed of weather.
A neural nowcasting core ingests radar, satellite, and surface telemetry every
sixty seconds — resolving storms before conventional models finish a single
integration step.
From the runway to the farm row, Itheons tunes its forecast resolution
and alerting cadence to the people who rely on weather to make decisions.
Aviation
Micro-turbulence mapping and cross-wind forecasts at field-by-field granularity for general aviation and drone operators.
Agriculture
Soil-temperature models, frost alerts, and irrigation scheduling tuned to crop stage and local topography.
Outdoor recreation
Trail-level precipitation windows, lightning proximity, and UV exposure tuned to elevation and canopy.
Emergency response
Severe-weather nowcasts with 45-minute lead time, push to incident command, and automated shelter-in-place triggers.
Headquartered in Welch, Oklahoma — with edge nodes across North America and global expansion underway.
How it works
Three stages. Sixty seconds.
A single forecast cycle ingests raw atmosphere, resolves it on our neural core,
and delivers targeted outputs to the devices that act on them.
01
Ingest
Radar sweeps, satellite channels, METAR stations, and IoT telemetry stream into the regional mesh every 60 seconds.
02
Resolve
A neural nowcasting core integrates the fields at 0.25km grid resolution — resolving cells before conventional models finish one step.
03
Deliver
Tailored outputs — push alerts, MQTT topics, GRIB2 payloads — reach your devices with a sub-second edge hop.
Field reports
Trusted by operators who can't be wrong.
"The 45-minute precipitation window saved us a full crew relocation during a spring dispatch. The old NWS model gave us 15."
Marcus TellerIncident Commander, Rural EMS · Kansas
"Frost alerts tuned to my actual elevation. I stopped losing crews to false warnings from the regional forecast."
Lena ParkOrchard operations · Willamette Valley
"I fly bush ops. Micro-turbulence resolution at the field scale is the first thing I check before wheels-up."
D. AlvarezPart 135 Captain · Alaska
FAQ
Questions, resolved.
How is Itheons different from the National Weather Service forecast?
The NWS runs models at 3km resolution on a multi-hour cycle. Itheons resolves
the same fields at 0.25km on a 60-second cycle — a 144× improvement in spatial
density and a 720× improvement in update cadence.
Do you cover areas outside the United States?
North America is fully online today. EMEA and APAC edge nodes deploy through
the end of the year. Contact us for a coverage map at your specific coordinates.
Can I integrate Itheons with my existing systems?
Yes. Outputs are available via REST, GraphQL, MQTT, and native GRIB2 — with
prebuilt connectors for common dispatch, SCADA, and incident-management platforms.
What's the accuracy on severe-weather lead time?
98.4% accuracy at the 6-hour mark on peer-reviewed benchmarks. Severe convective
nowcasts currently average a 45-minute lead time, with 2024 benchmarks pushing past 60.
How do you handle data privacy?
Telemetry is aggregated at the edge; no personal location data leaves the device
without explicit opt-in. Enterprise deployments support on-prem ingest and
SOC 2 Type II attestation.
Get Itheons
Your atmosphere. On every device.
Download the Itheons app, plug into the API, or talk to our team about
an enterprise deployment. It takes under five minutes to get your first forecast.