CVAT (Computer Vision Annotation Tool) is an open-source data annotation platform for images, video, and 3D point clouds, originally developed at Intel and now maintained by CVAT.ai. It supports bounding boxes, polygons, keypoints, segmentation, and object tracking, with AI-assisted labeling to speed up work. Teams can self-host the MIT-licensed Community edition for free or use the hosted CVAT Online service. Paid plans start at $23 per month billed annually for Solo, with per-seat Team plans and custom Enterprise pricing.
open-sourcecomputer-visionvideo-annotationself-hosted
Overview
- Founded
- 2018
- Pricing tier
- High
- Startup-friendly
- Yes
- Enterprise-ready
- Yes
Data labeling
- Managed workforce
- No
- Model-assisted labeling
- Yes
- Video annotation
- Yes
- Text annotation
- No
- Open-source
- Yes
- Self-hostable
- Yes
- Free tier
- Yes: Free self-hosted Community edition (MIT) and a free online plan; Solo from $23/mo billed annually.