PyOD
pyod.readthedocs.io
- Free plan
- Windows
- macOS
- Linux
- API
- Self-hosted
What PyOD 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 choose the classic detector interface or ADEngine, which profiles data, selects benchmark-backed detectors, runs detectors in parallel, calculates consensus scores, and reports diagnostics. ADEngine also works as a standalone Python API without an LLM. PyOD offers an agentic investigation workflow and activation paths for Claude Code, Codex, and MCP-compatible agents. Installation is documented through pip, conda-forge, or source, and requires Python 3.9 or higher. Optional pip extras provide capabilities such as PyTorch detectors, graph detectors, embeddings, audio, and an MCP server. ADEngine characterizes its quality verdict as a heuristic rather than a guarantee and recommends validation against held-out labels or domain review. The project says it serves academic research and commercial products.
Who it suits
PyOD suits Python users working on anomaly detection across the listed data types, including researchers and commercial product teams. ADEngine's agent-related options may also interest users building agent workflows.
Good to know, and what to watch
- Documents 61 detectors across six data types.
- ADEngine can run as a standalone Python API.
- Install through pip, conda-forge, or source.
- Free library with a common detector API.
- Requires Python 3.9 or higher.
- Some capabilities require optional pip extras.
- ADEngine's quality verdict is a heuristic, not a guarantee.
- Loading pickle or joblib artifacts requires trusted input.
Plans
Every plan and pricePyOD
Free plan
Getting started
- Open the PyOD documentation website
- Install with pip, conda-forge, or from source
- Use Python 3.9 or higher
- Choose the classic detector API or ADEngine
- Add optional pip extras for required detectors or integrations
- Activate the od-expert skill or optional MCP server when using supported agents
Limits to check
PyOD has no paid tiers or free-trial limits. Optional capabilities depend on installing the relevant pip extras, and ADEngine’s quality assessment remains heuristic.
The verdict
PyOD offers a broad detector catalog and both classic and orchestrated workflows. Treat ADEngine's assessment as a starting point and validate results with held-out labels or domain review.
Questions people ask
Is PyOD free?
Yes. The PyOD plan is free at no cost, and it is an open-source Python library.
Which data types does PyOD support?
It documents detectors for tabular, time-series, graph, text, image, and audio data.
What installation methods are available?
You can install PyOD through pip, conda-forge, or from source. Python 3.9 or higher is required.
Can PyOD work with agents?
Yes. It provides activation paths for Claude Code, Codex, and MCP-compatible agents, including an od-expert skill and optional MCP server.
Does ADEngine require an LLM?
No. ADEngine can be used as a standalone Python API without an LLM.
What the maker publishes
- Agent integrations
- The installation guide describes activation paths for Claude Code, Codex, and MCP-compatible agents. source
- Agent support
- PyOD provides an od-expert skill for Claude Code and Codex, plus an optional MCP server for MCP-compatible agents. source
- Contribution criterion
- PyOD says contributors to newly proposed detectors should commit to at least two years of maintenance. source
- Data types
- PyOD 3 documents detectors for tabular, time-series, graph, text, image, and audio data. source
- Detector catalog
- The documentation describes 61 detectors across multiple data types, exposed through one API. source
- Detector count
- The documentation lists 61 detectors across its supported data types. source
- Distribution
- The guide documents installation through pip, conda-forge, or from source. source
- Founded
- 2017 source
- Install options
- The package is distributed through pip and conda-forge and can also be installed from source. source
- Integrations
- Optional pip extras enable PyTorch, SUOD, XGBoost, model combination, thresholding, embeddings, OpenAI embeddings, Hugging Face encoders, graph models, MCP, and audio features. source
- Intended users
- The project says PyOD serves academic research and commercial products worldwide. source
- Lifecycle orchestration
- ADEngine profiles data, selects benchmark-backed detectors, runs multiple detectors in parallel, computes consensus scores, and reports diagnostics. source
- Optional components
- Optional pip extras include support for PyTorch detectors, graph detectors, embeddings, audio, and an MCP server. source
- Project history
- The About page says Dr. Yue Zhao initialized the project in 2017. source
- Purpose
- PyOD is a Python library for anomaly detection. source
- Python integration
- ADEngine can be used as a standalone Python API without an LLM. source
- Requirements
- The installation guide lists Python 3.9 or higher as a requirement. source
- 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. source
- Runtime requirement
- The installation guide requires Python 3.9 or higher. source
- 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. source
- Support
- The FAQ invites users to open an issue or contact the maintainer at [email protected]. source
- 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. source
- Usage
- PyOD offers a classic detector API, ADEngine lifecycle orchestration, and an agentic investigation workflow. source
Checked against the maker's pages on 30 Sept 2026.