AI for Time Series Intelligence
Zero-shot forecasting and anomaly detection with enterprise-level performance. No training. No onboarding. No infrastructure. Just results.
Why GradientCast?
Forecasting and anomaly detection are complex problems that typically require specialized teams. GradientCast uses transformer-based models—the same technology powering LLMs—trained specifically for time series forecasting and anomaly detection, delivering exceptional 0-shot performance. Complex AI, made accessible.
Zero-Shot Learning
Our foundational models have learned patterns from billions of time series. No training data needed. No labeling. No waiting. Just send your data and get predictions.
API-First Design
pip install, import, forecast. That's it. Get production-ready predictions in 3 lines of Python. Works with pandas. Full type hints. Comprehensive error handling.
Enterprise Ready
GPU-accelerated inference on Azure infrastructure. Auto-scaling from prototype to production. 99.9% uptime SLA. SOC 2 compliant data handling.
Our Products
Three specialized AI services for different time series challenges
GradientCastFM
State-of-the-art 0-shot forecasting powered by transformer-based foundational models. Predict future values across any domain.
- Multi-series forecasting
- Covariate support
- All frequencies (hourly to yearly)
PulseAD
Forecast-based deviation detection. Uses FM to predict expected values and flags significant deviations in real-time.
- Configurable thresholds
- Per-dimension overrides
- Multi-metric monitoring
DenseAD
Intelligent pattern-based anomaly detection with severity classification. Learns normal behavior and identifies subtle deviations.
- Severity classification
- Pattern learning
- Noise filtering
Production-ready in minutes
Our SDK is designed for developers. Install with pip, import the client, and start forecasting. Full type hints, pandas support, and comprehensive error handling included.
Install the SDK
pip install gradientcast
Initialize the client
Pass your API key
Call the API
Get predictions instantly
from gradientcast import GradientCastFM
# Initialize client
fm = GradientCastFM(api_key="your-api-key")
# Get forecast
result = fm.forecast(
input_data={"sales": [100, 120, 115, 130]},
horizon_len=7,
freq="D"
)
print(result["sales"])
# [142.5, 148.2, 151.8, 155.3, ...]Built for Every Industry
From e-commerce to infrastructure monitoring, GradientCast powers time series intelligence across diverse domains.
E-commerce
FMDemand forecasting, inventory optimization
Finance
FMRevenue projection, cash flow prediction
SaaS
FMUser growth forecasting, churn prediction
Infrastructure
PulseADTraffic anomaly detection, outage prediction
IoT
DenseADSensor anomaly detection, predictive maintenance
Security
DenseADBehavioral anomaly detection, threat identification
Simple, Transparent Pricing
Start free. Scale as you grow. No hidden fees.
Ready to transform your time series workflows?
Join teams using GradientCast to power their forecasting and anomaly detection.