CVAT.ai is the company that maintains CVAT, the Computer Vision Annotation Tool, a well-established open-source platform for labeling image, video, and 3D data. Originally developed at Intel and later stewarded within the OpenCV community, CVAT spun out into an independent company that now leads its development and offers commercial editions alongside the free project.
What they provide
CVAT supports the core tasks of supervised computer vision, including object detection, classification, tracking, and segmentation. Annotators can work with bounding boxes, polygons, masks, polylines, keypoints, cuboids, and tags across images, video, and 3D point clouds. The tool supports AI-assisted labeling by connecting detection, segmentation, and tracking models to speed up work, and provides quality assurance, team collaboration, analytics, and developer APIs. It is available in several forms: a free, self-hosted community edition; a hosted online edition with free and paid tiers; and an enterprise edition adding private hosting, SSO, and dedicated support. CVAT.ai also offers labeling services for teams that want annotation delivered as a service.
Where they fit
In the AI data supply chain, CVAT occupies the open-source tooling layer for computer vision, with a large community and broad adoption across research and production teams. Its long history and permissive availability make it a common default for teams that want a proven, self-hostable annotation tool without vendor lock-in. The hosted and enterprise editions extend that base for organizations that prefer managed infrastructure or need collaboration, security, and support at scale, positioning CVAT.ai between open-source software and a commercial annotation platform.
ImageVideo