CVAT helps teams turn images, video, 3D data, and audio into labeled datasets for AI workflows. Its visual tools cover boxes, pixel masks, polygons, keypoints, skeletons, and object tracks. Teams can review work through validation and acceptance stages, and assess quality with ground truth, honeypots, and consensus workflows. Storage connections include Amazon S3, Azure Blob Storage, Google Cloud Storage, S3-compatible buckets, and self-hosted storage. CVAT names SAM 2, SAM 3, Ultralytics, Hugging Face models, and custom models as ways to assist labeling. It exports to more than 20 formats, including COCO, YOLO, KITTI, Cityscapes, and Pascal VOC; automation is available through its API, SDK, and CLI. Deployment options include managed cloud and self-hosting, including on-premises and air-gapped deployments. A free plan is available, while online paid plans start at $23/user/mo billed annually. CVAT lists encryption, role-based access, SSO, logging, and audit trails among its security controls, but says it does not currently hold ISO 27001 or SOC 2 certifications.
Who it is for
CVAT suits AI teams labeling visual, 3D, or audio data, including teams that need review workflows and dataset exports. Managed cloud and self-hosted options serve different deployment needs.
What is good
- Supports image, video, 3D, and audio labeling
- Offers validation and acceptance review stages
- Exports to more than 20 formats
- Available as managed cloud or self-hosted
What to know first
- Community lacks SSO and automated quality assurance
- Community lacks SAM and built-in AI agent integrations
- Does not currently hold ISO 27001 or SOC 2
Verdict
CVAT combines varied annotation tools with review workflows, storage connections, and broad export options. Its Community edition omits several listed Enterprise capabilities, and the company says it lacks ISO 27001 and SOC 2 certifications.
CVAT plans and pricing
All plansCompared on video annotation software
- Free plan
- Yes
- Model-assisted labeling
- Yes
- API access
- Yes
