roboflow/supervision

Health v2 · maintenance only

Adoption not part of the score

51063 stars · 4861 forks observed · 2026-09-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-09-21, confidence 0.97

Supervision is an open‑source Python library that provides reusable utilities for building computer‑vision applications, from dataset handling to real‑time annotation and evaluation. It works with any detection, classification or segmentation model and offers tools for drawing, tracking, zoning and benchmarking detections.

Use cases

• draw bounding boxes and masks on images or video streams
• track objects across video frames with persistent IDs
• convert and load datasets between COCO, YOLO and Pascal VOC formats
• benchmark model performance using mAP and confusion matrices
• count detections inside custom polygon zones in real time
• integrate any CV model (YOLO, Detectron2, Transformers, etc.) for inference

When to choose

• You need a lightweight, model‑agnostic toolkit for quick CV prototyping in Python.
• You want built‑in utilities for annotation, tracking, and zone‑based counting without writing custom code.
• Your workflow requires easy dataset format conversion and performance evaluation.

When to avoid

• You require a full‑stack deep‑learning framework with training pipelines (e.g., PyTorch, TensorFlow).
• You need low‑level GPU kernels or custom C++ extensions for maximum performance.
• Your project is not Python‑based or targets non‑Python environments.

Facets

library · maturity active

image-processing video-processing computer-vision computer-vision deep-learning artificial-intelligence python windows cross-platform model-agnostic annotation-tools zone-counting map-evaluation linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="roboflow/supervision")