VGG Image Annotator

VIA, the VGG Image Annotator from Oxford's Visual Geometry Group, takes a deliberately minimal approach to dataset labeling: the entire tool is a single HTML file, typically under 400 KB, that runs fully offline in any modern browser with nothing to install. Annotators draw rectangles, circles, ellipses, polygons, and points on images and attach textual attributes to each region, then export the results for training or analysis. The project spans three generations: VIA 2 focuses on image annotation, while VIA 3 extends to video and audio with temporal segment marking and spatial region tracking across frames, and newer builds add shared collaborative projects. Created by Abhishek Dutta, Andrew Zisserman, and collaborators, it is open source under the BSD-2-Clause license, permitting commercial use, with code hosted on GitLab. Its zero-setup nature has made it a staple in academic computer vision labs and for quick labeling jobs where a full annotation platform would be overkill; the group reports millions of uses.