Skip to content
SPECIFICTION
Get listed

PyOD

pyod.readthedocs.io

Free plan WindowsmacOSLinuxAPISelf-hosted
Visit pyod.readthedocs.io
Spec score 8.4 #3 of 22 in Anomaly Detection Software

What it is

PyOD is a free Python library for anomaly detection, with documented detectors for tabular, time-series, graph, text, image, and audio data. Its documentation lists 61 detectors available through a common API. Users can work with the classic detector API or ADEngine, which profiles data, selects benchmark-backed detectors, runs multiple detectors in parallel, computes consensus scores, and reports diagnostics. ADEngine also works as a standalone Python API without an LLM. PyOD includes an agentic investigation workflow and activation paths for Claude Code, Codex, and MCP-compatible agents, including an od-expert skill and optional MCP server. Installation is available through pip, conda-forge, or source, and requires Python 3.9 or higher. Optional pip extras enable capabilities including PyTorch, graph detectors, embeddings, audio, and MCP support. The project serves academic research and commercial products. Its model persistence guidance warns that pickle and joblib can execute arbitrary Python code during deserialization, requiring callers to pass trusted=True when loading artifacts. ADEngine describes its quality verdict as a heuristic rather than a guarantee and recommends validation against held-out labels or domain review.

Spec sheet

Purpose
PyOD is a Python library for anomaly detection.Read 30 Sep 2026 · pyod.readthedocs.io
Intended users
The project says PyOD serves academic research and commercial products worldwide.Read 30 Sep 2026 · pyod.readthedocs.io
Integrations
Optional pip extras enable PyTorch, SUOD, XGBoost, model combination, thresholding, embeddings, OpenAI embeddings, Hugging Face encoders, graph models, MCP, and audio features.Read 30 Sep 2026 · pyod.readthedocs.io
Support
The FAQ invites users to open an issue or contact the maintainer at [email protected].Read 30 Sep 2026 · pyod.readthedocs.io
Founded
2017Read 28 Sep 2026 · pyod.readthedocs.io
Agent integrations
The installation guide describes activation paths for Claude Code, Codex, and MCP-compatible agents.Read 30 Sep 2026 · pyod.readthedocs.io
Agent support
PyOD provides an od-expert skill for Claude Code and Codex, plus an optional MCP server for MCP-compatible agents.Read 30 Sep 2026 · pyod.readthedocs.io
Contribution criterion
PyOD says contributors to newly proposed detectors should commit to at least two years of maintenance.Read 30 Sep 2026 · pyod.readthedocs.io
Data types
PyOD 3 documents detectors for tabular, time-series, graph, text, image, and audio data.Read 30 Sep 2026 · pyod.readthedocs.io
Detector catalog
The documentation describes 61 detectors across multiple data types, exposed through one API.Read 30 Sep 2026 · pyod.readthedocs.io
Detector count
The documentation lists 61 detectors across its supported data types.Read 30 Sep 2026 · pyod.readthedocs.io
Distribution
The guide documents installation through pip, conda-forge, or from source.Read 30 Sep 2026 · pyod.readthedocs.io
Install options
The package is distributed through pip and conda-forge and can also be installed from source.Read 30 Sep 2026 · pyod.readthedocs.io
Lifecycle orchestration
ADEngine profiles data, selects benchmark-backed detectors, runs multiple detectors in parallel, computes consensus scores, and reports diagnostics.Read 30 Sep 2026 · pyod.readthedocs.io
Optional components
Optional pip extras include support for PyTorch detectors, graph detectors, embeddings, audio, and an MCP server.Read 30 Sep 2026 · pyod.readthedocs.io
Project history
The About page says Dr. Yue Zhao initialized the project in 2017.Read 30 Sep 2026 · pyod.readthedocs.io
Python integration
ADEngine can be used as a standalone Python API without an LLM.Read 30 Sep 2026 · pyod.readthedocs.io
Requirements
The installation guide lists Python 3.9 or higher as a requirement.Read 30 Sep 2026 · pyod.readthedocs.io
Result quality limits
ADEngine describes its quality verdict as a heuristic, not a guarantee that results are correct, and recommends validation against held-out labels or domain review.Read 30 Sep 2026 · pyod.readthedocs.io
Runtime requirement
The installation guide requires Python 3.9 or higher.Read 30 Sep 2026 · pyod.readthedocs.io
Security guidance
The model persistence guide warns that pickle and joblib can deserialize arbitrary Python code and requires callers to pass trusted=True before loading artifacts.Read 30 Sep 2026 · pyod.readthedocs.io
Support and community
The documentation links to a GitHub repository for source installation and examples; it does not state a paid support plan on the pages reviewed.Read 30 Sep 2026 · pyod.readthedocs.io
Usage
PyOD offers a classic detector API, ADEngine lifecycle orchestration, and an agentic investigation workflow.Read 30 Sep 2026 · pyod.readthedocs.io

Plans and prices

PyOD Free Open-source Python library · optional capabilities require pip extras Read 30 Sep 2026 · pyod.readthedocs.io

Every PyOD plan and price →

Strong points and limits

Strong points

  • Documents 61 detectors across six data types.
  • Offers a common API and parallel ADEngine workflows.
  • Can use ADEngine without an LLM.
  • Install through pip, conda-forge, or source.
  • Optional extras add graph, audio, and embedding capabilities.

Limits

  • Requires Python 3.9 or higher.
  • Optional capabilities require pip extras.
  • ADEngine quality verdicts are heuristics, not guarantees.
  • Loading pickle or joblib artifacts can execute arbitrary code.

Who it suits

PyOD suits Python users in academic research or commercial products who need anomaly detection across different data types. ADEngine also offers orchestration for users seeking data profiling and combined detector results.

Verdict

PyOD provides a broad detector catalog and a more orchestrated ADEngine workflow, with multiple installation routes. Validate ADEngine findings against labels or domain review, and load persisted artifacts only when trusted.

Where Specifiction places it

Alternatives to PyODAll alternatives →