Nixtla/nixtla

TimeGPT-1: production ready pre-trained Time Series Foundation Model for forecasting and anomaly detection. Generative pretrained transformer for time series trained on over 100B data points. It's capable of accurately predicting various domains such as retail, electricity, finance, and IoT with just a few lines of code 🚀.

87
/ 100
Verified

Built on a transformer architecture trained on 100B+ time series datapoints, TimeGPT supports zero-shot inference for immediate predictions without fine-tuning, alongside optional fine-tuning with custom loss functions for domain-specific adaptation. The SDK integrates with pandas DataFrames and supports distributed computing frameworks (Spark, Dask, Ray), while native Snowflake deployment enables in-database forecasting via stored procedures and UDTFs without data movement. Key capabilities include exogenous variable incorporation, prediction intervals for uncertainty quantification, and handling of irregular timestamps.

3,792 stars and 123,695 monthly downloads. Actively maintained with 27 commits in the last 30 days. Available on PyPI.

Maintenance 23 / 25
Adoption 20 / 25
Maturity 25 / 25
Community 19 / 25

How are scores calculated?

Stars

3,792

Forks

317

Language

Jupyter Notebook

License

Last pushed

Mar 13, 2026

Monthly downloads

123,695

Commits (30d)

27

Dependencies

9

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/Nixtla/nixtla"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.