# sintel-dev/Orion

Unsupervised time series anomaly detection library

Repository: https://github.com/sintel-dev/Orion
Canonical: https://ross.abutalabs.com/products/orion
Homepage: https://sintel.dev/Orion/
Language: Python
License: MIT
License Family: permissive
Topics: anomaly-detection, deep-learning, machine-learning, time-series, benchmarking, signals, unsupervised-learning, orion, data-science, generative-adversarial-network
Last push: 2026-08-17T01:20:51+00:00

## Health v2 (maintenance only)
Score: 74/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 28, longevity 100
- inputs: {"age_days": 2962, "days_push": 17, "days_rel": 533, "gap_med": 81.5, "n_releases_24m": 3}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1368, forks 207 (observed 2026-08-28T04:04:31.682563+00:00)

## What it is
Orion is a Python machine learning library for unsupervised time series anomaly detection, offering verified ML pipelines that flag rare patterns for expert review. It includes benchmarking of methods and a backend database with REST API for building custom anomaly detection platforms.

## Use cases
- detect anomalies in time series data without labels
- find unusual patterns in satellite telemetry signals
- benchmark time series anomaly detection methods
- flag anomalous periods in sensor data for expert review
- detect collective anomalies in wind turbine signals
- build a custom time series anomaly detection platform

## When to choose
- you need unsupervised anomaly detection on time series with no labeled anomalies
- you want pre-verified ML pipelines instead of hand-tuning models
- you want to benchmark the latest anomaly detection research methods

## When to avoid
- you need supervised anomaly detection with labeled training data
- you need real-time streaming anomaly detection at low latency
- you need a production-stable tool - the project is pre-alpha

## Facets
- artifact type: library
- maturity: experimental
- function: machine-learning, deep-learning, benchmarking, data-science
- domain: machine-learning, time-series, data-science
- platform: python
- tags: anomaly-detection, time-series, unsupervised-learning, generative-adversarial-networks, signals, automl

## Member repositories
- sintel-dev/Orion (main) score 74

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:31.682563+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:41:04.249994+00:00, confidence not recorded.
  - readme: https://github.com/sintel-dev/Orion (fetched 2026-08-28T04:04:31.682563+00:00, sha eaf032a5b48a)
  - homepage: https://sintel.dev/Orion/ (fetched 2026-08-29T11:58:05.343345+00:00, sha 498930909cd5)
- Data as of 2026-08-30T08:39:29.467469+00:00.