- ABOD: cast non-float numeric input (integer lists and DataFrames, bool, float16) to float64 before the Numba angle kernel, which previously raised a TypingError. Scores for float32 and float64…
- XGBOD: keep estimatorlist and standardizationflaglist as given and store the resolved detectors in estimatorlist and standardizationflaglist, so getparams() and clone() round-trip. #767, fixes #755.
- Sampling and KPCA: store randomstate as given and resolve it in fit, so repeated fits with an integer seed are reproducible. #753, fixes #750.
- INNE: call BaseDetector.init, so an out-of-range contamination raises ValueError like every other detector. #759, fixes #752.
- HBOS and LODA: getoptimalnbins accepts 1 to 3 training samples and rejects empty input with a clear message. #763, fixes #762.
- k-Shape: negate the SBD lag when aligning cluster members, so refinement moves members onto the centroid. #734.
Pyod by Yue Zhao
A Python library for anomaly detection across tabular, time series, graph, text, image, and audio data. 60+ detectors, benchmark-backed ADEngine orchestration, and an agentic workflow for AI agents.
AI data analysis agents
New · Added to AgentGid on Oct 15, 2026Open source · BSD-2-ClauseFree to start
Recently added on Oct 15, 2026. The description, category and pricing were filled in from public metadata and have not been checked by an editor yet.
Price
Free (OSS)
free and open source
Popularity
#1 in Data
Gid Score 49 / 100 · #44 of 260 ranked AI agents
Trend
Slowing down
usage down 9.2% over 30 days
Development
Updated recently
latest 3.6.7, 9d ago
Community
10.1K GitHub stars
690K package downloads a week
Usage data: latest readings Oct 19 to Oct 21, 2026, last checked Oct 21, 06:33 UTC. Date of each number
What Pyod does
It is free and open source (BSD-2-Clause); you pay only for the models it calls.
Key facts
| Maker | Yue Zhao |
|---|---|
| First released | October 2017 |
| License | BSD-2-Clause |
| Main language | Python |
| Open issues and PRs | 223 |
Pricing
Open source under the BSD-2-Clause licence; check the website for any paid hosted plans.
, checked Oct 15, 2026. Prices change often; confirm before you buy.
Quick answers
- Free to start?Yes
- Open source?Yes (BSD-2-Clause)
- Closest alternativeHex · all alternatives
AgentGid articles about Pyod
Get started with Pyod
| PyPI | pip install pyod |
|---|
These commands use the packages tracked on this page. The vendor may recommend a different installer.
Pyod alternatives
Best Pyod alternatives compared · All AI data analysis agents
| # | Agent | Gid Score | Downloads 7d | 7d | 30d | Stars | Latest release | Price | |||
|---|---|---|---|---|---|---|---|---|---|---|---|
| 133 | 33 | — | — | — | — | — | Free + $36/mo | ||||
| 141 |
PandasAISinaptik AI |
32 | 8,784 | up 7.2% | — | 23.9K+4/day | 3.0.012mo ago | Free | |||
| 189 |
Julius AIJulius AI |
23 | — | — | — | — | — | Free + $20/mo | |||
| 201 |
WrenAICanner |
20 | — | — | — | 17.9K+10/day | 0.3.1yesterday | Free + $179/mo | |||
| 251 |
Powerdrill BloomPowerdrill |
6 | — | — | — | — | — | Free + $16.58/mo | |||
| — |
Databricks GenieDatabricks |
— | — | — | — | — | — | Usage-based | |||
| No agents match that filter. | |||||||||||
Head to head: Pyod vs Hex · Pyod vs PandasAI · Pyod vs Julius AI · Pyod vs WrenAI
Tools used with Pyod
What 32 public GitHub projects that depend on Pyod use with it, from their dependency files.
- NumPy100.0%
- scikit-learn93.8%
- pandas90.6%
- Matplotlib87.5%
- SciPy75.0%
- tqdm59.4%
- joblib59.4%
- Requests59.4%
- Plotly56.2%
- statsmodels56.2%
Used far more often here than in AI-built projects overall: what sets Pyod projects apart.
- pmdarima31.2% · 1940.9×
- Cython37.5% · 582.3×
- parso37.5% · 332.7×
- cloudpickle40.6% · 315.4×
- Jedi34.4% · 213.5×
- contourpy31.2% · 194.1×
- statsmodels56.2% · 183.9×
- Numba43.8% · 181.2×
Ratings from people who use Pyod
No ratings yet. Used Pyod? Rate it and say what it is good and bad at.
Pyod news
-
10 agents added to AgentGid
Pyod, Rakazo, OpenExecutive, Skill Recorder, Fuxi, Open Science, Casdoor, Tutti, Zvec Grep, Financial-API
-
Pyod 3.6.7 released
ABOD: cast non-float numeric input (integer lists and DataFrames, bool, float16) to float64 before the Numba angle kernel, which previously raised a TypingError. Scores for float32 and float64…; XGBOD: keep estimatorlist and standardizationflaglist as given…
What changed: recent Pyod releases
5 stable releases in the last 90 days · Full changelog
- LOCI: preserve the constructor parameter k for cloning and use contamination to derive the fitted threshold and binary labels. k still controls the radius-scan early stop. #707, fixes #194.
- Pearson correlation: compute all row pairs in the unweighted pearsonrmat utility when there are more rows than features; previously some correlations incorrectly remained 1. #731.
- Deep-learning base class: return the fitted estimator from BaseDeepLearningDetector.fit. #740.
- VAE: preserve logvarclip as supplied to the constructor so sklearn.clone works, while validating the training value separately. #738.
- Sampling: retain a fractional subsetsize across fits instead of replacing the constructor argument with the first computed sample count. #737.
- Kernel PCA: forward the requested ncomponents to the underlying estimator. #730.
Earlier releases
- Generated data changes for the 46 degenerate randomstate values listed above. All other seeds are bit-identical.
- ROD decisionscores and labels change for any input containing a point at the geometric median. Scores on data with no such point are bit-identical.
- CBLOF emits a new FutureWarning for a non-default njobs. It repeats on each fit call rather than once per process, because scikit-learn's internal use of catchwarnings invalidates the caller's…
- generatedata no longer accepts offsets that were never scalars but which NumPy used to coerce, such as '2', b'2', and one-element sequences or arrays. Those forms were outside the documented…
- The README quick start called clf.fit(Xtrain) without ever defining Xtrain, and used visualize without importing it. Pasting it raised NameError. This block is the GitHub landing page and the PyPI…
- docs/install.rst documented a pytorch extra that does not exist. pip treats an unknown extra as a warning, so pip install pyod[pytorch] succeeded while installing none of the PyTorch stack, and…
- The fitted center is initialized from the data, detached, and stored in the new c attribute that the objective and scorer actually read.
- The optimizer was built with no learning rate (Adam's 1e-3 default) and passed l2regularizer=0.1 as weight decay, five orders of magnitude above the reference implementation's 5e-7. That decay…
- The best-epoch snapshot was stored but never loaded, so scoring used final-epoch weights.
- Local encoder paths fall back to the HuggingFace backend when sentence-transformers is absent (#701,
- testresolvestinstancenodownload no longer breaks on sentence-transformers 5.6, which rejects the empty modules=[] construction the test relied on (#710).
- save/load round-trip coverage now spans 23 detectors, up from 2, and 21 of them gain clone assertions for unfitted equivalence and refit score reproduction (#708,
- Persistence guide, decision tree, troubleshooting table, and the save/load example updated to pass trusted=True.
- Regression tests for both loaders' guard and the internal trust forwarding.
Pyod popularity and usage data
Pyod was downloaded 690K times from PyPI in the week ending Oct 19, 2026, up 12.5% on the week before and down 9.2% over the last 30 days compared with the 30 before. The GitHub repository has 10.1K stars, 54 of them added in the last 30 days. The latest stable release is 3.6.7, published Oct 12, 2026; there were 1 stable releases in the past 30 days.
Download the raw daily series: pyod.csv
Gid Score breakdown
Adoption63
Community42
Attention—
Momentum41
Each component is scored 0 to 100 from public signals; a dash means no data for it. Methodology.
How fresh these numbers are
| Number | Latest reading | Last checked |
|---|---|---|
| Package downloads | Oct 19, 2026 | Oct 21, 06:31 UTC |
| GitHub stars | Oct 21, 2026 | Oct 21, 06:33 UTC |
We fetch every source once a day. Registries publish downloads and page views a day or two late, so their latest reading trails the check. A reading more than 4 days old is marked out of date; the Gid Score still uses it until a newer one arrives.
Pyod FAQ
How much does Pyod cost?
Open source under the BSD-2-Clause licence; check the website for any paid hosted plans.
Is Pyod open source?
Yes. Pyod is open source, released under the BSD-2-Clause license.
How popular is Pyod?
Pyod was downloaded 690K times from PyPI in the week ending Oct 19, 2026, up 12.5% on the week before and down 9.2% over the last 30 days compared with the 30 before. The GitHub repository has 10.1K stars, 54 of them added in the last 30 days. The latest stable release is 3.6.7, published Oct 12, 2026; there were 1 stable releases in the past 30 days.
What is the latest version of Pyod?
The latest stable release we track is 3.6.7, published on Oct 12, 2026.
What are the alternatives to Pyod?
The closest alternatives in the same category by Gid Score are Hex, PandasAI, Julius AI, WrenAI.
For the makers of Pyod badges, corrections
Add a live badge to your README or site:
[](https://agentgid.com/agents/pyod/)[](https://agentgid.com/agents/pyod/)[](https://agentgid.com/agents/pyod/)Something wrong or outdated on this page? Send a correction with a link to the official source. More for makers.
Where these numbers come from
PyPI: pyod. GitHub: yzhao062/pyod. Site rank: pyod.dev on the Tranco list. See data sources for how each one is collected.
Product facts and pricing were checked against: pyod.dev, github.com, pypi.org.