VoTT

VoTT (Visual Object Tagging Tool) is an open-source image and video annotation application developed by Microsoft to streamline dataset creation for computer vision and machine learning projects. It provides an intuitive interface for labeling objects with bounding boxes, polygons, or tags, making it easy to prepare high-quality datasets for AI model training.

Key Features of VoTT:

• Simple and interactive labeling of images and videos using bounding boxes, polygons, or region-based tagging.
• Seamless export to popular machine learning dataset formats such as YOLO, COCO, Pascal VOC, and TensorFlow TFRecord.
• Supports multiple storage backends including local file systems and Azure Blob Storage for large-scale dataset management.
• Project-based workflow for organizing datasets, tags, and labeling progress across teams.
• Cross-platform compatibility — runs on Linux, Windows, and macOS via Electron or can be deployed as a self-hosted web app.
• Fully open-source and extensible, allowing customization, automation, and integration into MLOps pipelines.
• Lightweight design focused on accuracy, speed, and usability for computer vision model preparation.

To check your VoTT version:

# vott --version

Disclaimer: VoTT is open-source software originally developed by Microsoft and distributed under the MIT License. It is provided “as is,” without warranty of any kind. Users are responsible for ensuring the security, integrity, and compliance of their datasets. Community versions and builds may differ from the archived Microsoft release; always verify sources before use in production environments.