Infrastructure-grade forecasting from weather model to EMS output.

Gridvynt ingests NWP ensemble runs, corrects them at the asset level using calibrated ML, and publishes probabilistic forecasts that your reserve commitment and scheduling workflows can act on directly.

Request Pilot Access Forecasting Engine details

Core modules

Four integrated modules. One deployable system.

Forecasting Engine

Four-member NWP ensemble (GFS, ECMWF-IFS, NAM, HRRR) with site-level ML correction. 15-minute intervals, 72-hour horizon, P10/P50/P90 probabilistic output.

Explore

Grid Balancing

Translates forecast probability bands into deterministic thermal reserve targets. Integrates NERC BAL-001/002 margin logic.

Explore

Data Infrastructure

Time-series ingestion pipeline for SCADA, historian, and NWP sources. 99.7% uptime SLA. Sub-5-minute data latency.

Explore

Platform specs

The numbers behind the platform.

15 min

Forecast interval at full 72-hour horizon

NWP ensemble members (GFS, ECMWF, NAM, HRRR)

Model recalibration cycle against actuals

gridvynt — forecast API · v2

$ curl -s "https://api.gridvynt.com/v2/forecast?asset=SAR-001&horizon=72h&interval=15min" \
  -H "Authorization: Bearer $GVT_API_KEY"

{
  "asset": "SAR-001",
  "horizon_h": 72,
  "interval_min": 15,
  "bands": ["p10", "p50", "p90"],
  "units": "MW",
  "points": [
    {"ts": "2026-06-14T00:00Z", "p10": 0.0, "p50": 0.0, "p90": 0.0},
    {"ts": "2026-06-14T00:15Z", "p10": 0.0, "p50": 0.0, "p90": 0.0},
    {"ts": "2026-06-14T06:00Z", "p10": 41.2, "p50": 52.8, "p90": 64.1},
    ... 284 more points ...
  ]
}

Ready to see the platform on your data?

We stand up a 30-day pilot on your actual asset register and SCADA feeds. No WRF license, no new hardware.

Request Pilot Access View Methodology