PySAD is a free, open-source Python framework for detecting anomalies in streaming univariate and multivariate data. Its online models update as each new data instance arrives, supporting sequential detection. The project describes 16 online detectors, including xStream, LODA, RS-Hash, Half-Space Trees, and Robust Random Cut Forest. It also includes stream simulators, evaluators, preprocessors, statistic trackers, postprocessors, and probability calibrators. PySAD connects batch anomaly detectors from PyOD to streaming use, and supports supervised, semi-supervised, and unsupervised experiments. The documentation notes that streaming methods may retain a single instance or a small recent window to stay within memory and processing constraints. Install it with pip or from its GitHub source; the current README lists Python 3.10 or newer on Linux, macOS, and Windows. PySAD is self-hosted and distributed under a BSD 3-Clause license. The project says 17 labelled benchmark datasets download on first use.
Who it is for
PySAD suits developers and data practitioners building streaming anomaly detection workflows in Python. It supports experiments across univariate and multivariate data and supervised, semi-supervised, or unsupervised settings.
What is good
- Online models update with each new data instance.
- Describes 16 online detectors.
- Includes evaluation and preprocessing tools.
- Supports PyOD batch detectors in streaming settings.
- BSD 3-Clause license.
What to know first
- Requires Python 3.10 or newer, per the current README.
- Self-hosted deployment.
Verdict
PySAD brings online detectors and workflow utilities together for streaming anomaly detection. Its self-hosted, Python-based setup is suited to users who want to build and run their own detection workflows.
PySAD plans and pricing
All plansCompared on anomaly detection software
- Free plan
- Yes
- Detection method
- hybrid
- Real-time detection
- Yes
- Supported data
- univariate data; multivariate data; streaming data
- Deployment options
- self-hosted
