OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Nixtla

Nixtla provides TimeGPT-2.1, a pretrained generative time-series model for forecasting values and detecting anomalies, accessed through a Python SDK and other documented API language options.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Nixtla — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Nixtla when a project needs forecasts or anomaly detection from time-series data and a pretrained model workflow fits the task. The repository provides a Python example for electricity demand and describes a separate Snowflake deployment route.

Who it is for

  • Data analysts working with historical time-series data.
  • Developers building forecasting or anomaly-detection workflows.
  • Snowflake users seeking to run documented forecasting and anomaly components within their database environment.

Documented uses

  • Forecast future values from historical time-series data.
  • Request prediction intervals alongside a forecast.
  • Identify anomalies in historical observations or use the documented online detection path.
  • Set up forecasting and anomaly-detection components in Snowflake.
Before you choose

Requirements & limits

The documented Python SDK workflow requires an API key. The repository points to a free-trial page for obtaining one.

  • The Python quick start uses the Nixtla SDK and an API key.
  • The example requests a 24-hour forecast and 80% and 90% intervals.
  • The README warns that detect_anomalies() is removed in Nixtla 1.0 and identifies detect_anomalies_online() as the path requiring h and detection_size.
  • The Snowflake route uses a separate installation script and requires setup of external-access integrations and an API key.
  • The repository describes zero-shot inference, fine-tuning, multiple-series forecasting, exogenous variables and irregular timestamps as capabilities.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Request an electricity-demand forecast with the Python SDK

Use the documented NixtlaClient quick-start flow to request a 24-hour forecast with prediction intervals.

Before you start

  • Python environment with the Nixtla SDK and pandas installed.
  • A Nixtla API key.
  • Historical time-series data in the example’s expected format.

Steps

  1. Install the SDK with the command below and import pandas and NixtlaClient.

    Documented command
    pip install nixtla>=0.8.0

  2. Create a NixtlaClient using the API key.

  3. Load the documented electricity-demand CSV into a pandas DataFrame.

  4. Call forecast with h=24 and level=[80, 90]; optionally plot the forecast.

Expected outcome

A forecast for the next 24 hours with 80% and 90% prediction intervals; plotting is shown as optional.

Guide limitations

  • The guide follows the README’s electricity example and does not establish performance on other datasets.
  • The API key is required for the documented client workflow.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
16 October 2026
Original website title
Nixtla/nixtla · Official homepage title · retained in its source language
How we review sourcesSuggest a correction