CVAT
CVAT (Computer Vision Annotation Tool) is an open-source platform for annotating images, video, and 3D point clouds for vision AI. The CVAT REST API exposes projects, tasks, jobs, annotations, labels, organizations, memberships, and cloud storage integrations, available both self-hosted (MIT-licensed) and as the hosted CVAT Online service at app.cvat.ai.
APIs
CVAT Projects API
Create, list, retrieve, update, and delete annotation projects, the top-level container that groups tasks and shares a label set, plus dataset import and export operations.
CVAT Tasks API
Manage annotation tasks - create tasks, upload media data, retrieve metadata, import and export annotations and datasets, and run background data processing within a project.
CVAT Jobs API
Work with jobs, the assignable annotation units that subdivide a task - list, retrieve, update stage and state, manage assignees, and import or export job-level annotations.
CVAT Annotations API
Read, replace, and remove annotation data (shapes, tracks, and tags) on tasks and jobs, and trigger background import and export of annotations in supported dataset formats.
CVAT Labels API
List, retrieve, update, and delete labels and their attributes and sublabels, defining the annotation taxonomy shared by projects, tasks, and jobs.
CVAT Organizations and Memberships API
Manage organizations and their memberships - create and configure organizations, invite and manage members, assign roles, and scope projects, tasks, and storage to an organizati...
CVAT Cloud Storages API
Register and manage connections to external object storage (AWS S3, Azure Blob, Google Cloud Storage) used as data sources for tasks, including listing bucket content and previews.