# roboflow/trackers

Trackers gives you clean, modular re-implementations of leading multi-object tracking algorithms released under the permissive Apache 2.0 license. You combine them with any detection model you already use.

Repository: https://github.com/roboflow/trackers
Canonical: https://ross.abutalabs.com/products/trackers
Homepage: https://trackers.roboflow.com/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: multi-object-tracking, sort, bytetrack, oc-sort, bot-sort, dancetrack, mot17, soccernet, sportsmot, trackeval
Last push: 2026-08-26T15:26:30+00:00

## Health v2 (maintenance only)
Score: 85/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 96, longevity 36
- inputs: {"age_days": 506, "days_push": 7, "days_rel": 27, "gap_med": 26.5, "n_releases_24m": 7}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3717, forks 400 (observed 2026-08-28T04:08:15.382871+00:00)

## What it is
A Python library of clean-room, Apache 2.0 implementations of multi-object tracking algorithms including SORT, ByteTrack, OC-SORT, BoT-SORT, C-BIoU, and McByte. It is detector-agnostic, works natively with supervision.Detections, and includes Optuna-based hyperparameter tuning.

## Use cases
- track objects across video frames with any detection model
- compare multi-object tracking algorithms like ByteTrack and BoT-SORT
- keep stable track IDs when the camera is moving
- tune tracker hyperparameters for my scene and detector
- add object tracking to a YOLO or DETR pipeline without glue code
- benchmark trackers on MOT17 or SportsMOT

## When to choose
- you need permissively licensed (Apache 2.0) tracking code for closed-source products
- you already use the supervision ecosystem or any bounding-box detector
- you want readable, modifiable implementations rather than vendored wrappers
- you need camera motion compensation for moving cameras

## When to avoid
- you need full perception pipelines including detection models - this only tracks given detections
- you require AGPL-licensed ecosystem features from alternatives like BoxMOT
- you need tracking in a non-Python language

## Facets
- artifact type: library
- maturity: active
- function: computer-vision, machine-learning, benchmarking
- domain: computer-vision, machine-learning
- platform: python, cross-platform
- tags: multi-object-tracking, sort, bytetrack, oc-sort, bot-sort, object-detection, supervision, optuna, video

## Member repositories
- roboflow/trackers (main) score 85

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:15.382871+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 2026-08-29T18:29:22.473061+00:00, confidence not recorded.
  - readme: https://github.com/roboflow/trackers (fetched 2026-08-28T04:08:15.382871+00:00, sha ccc730687070)
  - homepage: https://trackers.roboflow.com/ (fetched 2026-08-29T09:24:09.781816+00:00, sha 36c6c3c2e4f9)
- Data as of 2026-08-30T08:39:29.467469+00:00.