Python Library for Anomaly Detection
9.8k 2026-04-18
yzhao062/pyod
A comprehensive Python library offering 60+ anomaly detectors for multi-modal data, featuring an agentic workflow for AI agents and benchmark-backed orchestration.
Core Features
Multi-modal anomaly detection across tabular, time series, graph, text, and image data.
Agentic workflow enabling AI agents to perform expert-level detection using plain English.
Full lifecycle support from raw data to explained anomalies and next-step guidance.
Over 60 built-in anomaly detection algorithms with a unified API.
Benchmark-backed routing (ADEngine) for optimal detector selection.
Quick Start
pip install pyodDetailed Introduction
PyOD 3 is a leading Python library designed for scalable and comprehensive anomaly detection. It provides a unified API for over 60 detectors capable of handling diverse data types including tabular, time series, graph, text, and image data. A key innovation is its agentic capability, allowing AI agents to conduct sophisticated anomaly investigations through natural language. PyOD supports the full anomaly detection lifecycle, from initial data processing to anomaly explanation and guidance, all while being backed by extensive benchmarks for performance optimization.