Retail & Operations
Use time series foundation models for demand forecasting. Predict sales volume, inventory needs, and capacity requirements through the TSFM.ai API.
Zero-shot predictionsMulti-SKU at scalePrediction intervals included
Use Cases
Forecasting use cases
Time series foundation models work across industries. Each page below covers a specific domain — the data shapes that matter, the right model choices, and how to go from a first API call to a production integration.
Energy & Utilities
Forecast grid load, renewable generation, and peak demand with time series foundation models. The TSFM.ai API handles the non-stationarity of energy data without per-asset model training.
Grid load predictionRenewable generationPeak demand estimation
Monitoring & Observability
Detect time series anomalies using foundation models. The TSFM.ai API provides prediction intervals and anomaly scoring that flag unusual behavior without labeled training data.
Prediction interval scoringZero labeled data neededAny time series domain
Industrial & Manufacturing
Predict equipment failures and sensor degradation with time series foundation models. The TSFM.ai API forecasts sensor readings and flags when predicted values approach failure thresholds.
Sensor forecastingFailure threshold estimationMulti-sensor support
Why foundation models work across use cases
Traditional forecasting requires a model trained per series or per domain. Foundation models generalize from pre-training on billions of observations — send any series and get forecasts without a fitting step. That means the same API call structure works whether you are forecasting retail demand, grid load, sensor degradation, or server CPU utilization.
Each use-case page covers the specific data shapes, evaluation criteria, and model recommendations for that domain. Start with the one closest to your workload.
Don’t see your use case?
The API handles any numerical time series. If your domain is not listed above, the forecasting API page covers the general integration pattern, and the playground lets you test on your own data before writing any code.