# mikel-brostrom/boxmot

BoxMOT: Pluggable Python and C++ SOTA multi-object tracking modules with support for axis-aligned and oriented bounding boxes

Repository: https://github.com/mikel-brostrom/boxmot
Canonical: https://ross.abutalabs.com/products/boxmot
Homepage: https://mikel-brostrom.github.io/boxmot/
Language: Python
License: AGPL-3.0
License Family: copyleft
Topics: strongsort, bytetrack, ocsort, deep-learning, segmentation, tensorrt, tracking-by-detection, yolo, botsort, deepocsort, multi-object-tracking, mot, mots, multi-object-tracking-segmentation, boosttrack, oriented-bounding-box-tracking, machine-learning
Last push: 2026-08-26T20:49:44+00:00

## Health v2 (maintenance only)
Score: 95/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 87, longevity 100
- inputs: {"age_days": 2259, "days_push": 7, "days_rel": 7, "gap_med": 3.0, "n_releases_24m": 75}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 8281, forks 1917 (observed 2026-08-28T04:10:20.058807+00:00)

## What it is
BoxMOT is a pluggable Python and C++ library providing state-of-the-art multi-object tracking (MOT) algorithms such as ByteTrack, BoT-SORT, StrongSORT, OC-SORT, DeepOCSORT, and BoostTrack, supporting both axis-aligned and oriented bounding boxes. It offers a unified CLI and Python API covering tracking, evaluation, tuning, ReID training, and export workflows with swappable detectors and trackers.

## Use cases
- track multiple objects in a video with yolo detections
- compare mot trackers on mot17 benchmark
- run bytetrack or botsort on my own detector output
- train and evaluate a person re-identification model
- track objects with oriented bounding boxes
- deploy a c++ multi-object tracker in production
- tune tracker hyperparameters for my dataset
- export reid embeddings for offline tracking

## When to choose
- you need state-of-the-art multi-object tracking with swappable trackers and detectors
- you want a single CLI/API for tracking, evaluation, tuning, and ReID workflows
- you need oriented bounding box tracking or a native C++ tracker backend

## When to avoid
- you only need single-object tracking
- you need a permissively licensed library (it is AGPL-3.0)
- you need real-time tracking on edge devices without GPU or TensorRT support

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, computer-vision, cli, sdk
- domain: computer-vision, machine-learning, deep-learning
- platform: python, cpp, cross-platform, cli
- tags: multi-object-tracking, tracking-by-detection, bytetrack, botsort, strongsort, ocsort, deepocsort, boosttrack, reid, yolo, tensorrt, oriented-bounding-boxes, mots, video, docker

## Member repositories
- mikel-brostrom/boxmot (main) score 95

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:20.058807+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 2026-08-29T17:28:06.689041+00:00, confidence not recorded.
  - readme: https://github.com/mikel-brostrom/boxmot (fetched 2026-08-28T04:10:20.058807+00:00, sha 9e6d5f48e80a)
  - homepage: https://mikel-brostrom.github.io/boxmot/ (fetched 2026-08-29T08:27:49.887664+00:00, sha afca54a1769b)
- Data as of 2026-08-30T08:39:29.467469+00:00.