CVAT

Open-source computer vision annotation platform originally developed by Intel, supporting image and video labeling for ML model training.

AI Training Computer Vision & Annotation

Visit CVAT

★ ★ ★ ★ ★ New (0 reviews)

Wilmington, USA HQ

11-50 company size

2018 founded

What is CVAT?

CVAT (Computer Vision Annotation Tool) is an open-source, web-based annotation platform originally developed by Intel in 2017 and released as open source in 2018. In 2022, CVAT spun out into an independent company, CVAT.ai, headquartered in Wilmington, Delaware with an EU office in Paphos, Cyprus. The platform is designed for annotating images and videos to create training datasets for computer vision models, supporting annotation types including bounding boxes, polygons, keypoints, image classification, object detection, object tracking, image segmentation, pose estimation, and 3D data. CVAT offers a free Community edition for self-hosting, plus commercial CVAT Online and Enterprise editions with additional features and dedicated support. The platform is widely used by researchers and engineering teams for building computer vision machine learning pipelines.

Mission & values

Be the industry-leading open-source data engine for machine learning — power image, video, and 3D annotation for 200,000+ developers worldwide.

Qualifications

For Data Annotation Specialist roles, expertise in data annotation workflows, an understanding of AI pipelines, and familiarity with the CVAT platform are beneficial. For technical roles such as Senior Python Developer, experience designing, developing, and documenting Python SDKs and APIs is required. Frontend Developer roles require proficiency in modern JavaScript frameworks. Copywriter positions need strong technical writing skills with an understanding of AI and computer vision concepts. No specific degree is universally required for annotation roles, but attention to detail and the ability to follow annotation guidelines accurately are essential. The company offers remote-worldwide positions as well as roles based in Cyprus. Experience with computer vision concepts (object detection, segmentation, tracking) is advantageous for all roles.

Leadership

Gary Bradski

President of OpenCV (CVAT now under OpenCV stewardship)

Renowned computer-vision researcher and OpenCV founder; endorsed CVAT as "the best data annotation tool around."

Funding

StageSeed

Investors

OpenCV (project stewardship)

Awards & recognition

200,000+ developers using CVAT globally

CVAT.ai
Industry-leading open-source CV annotation tool

OpenCV / Gary Bradski endorsement
Originated at Intel; transitioned to OpenCV stewardship

OpenCV

Frequently asked questions

Who built CVAT?

Originally developed by Intel for internal use, then open-sourced. Intel transferred CVAT's GitHub stewardship to OpenCV; the corporate entity is now CVAT.ai.

Is CVAT free?

The Community edition is free and open-source under the MIT license. The hosted SaaS has Free, Solo, Team, and Enterprise tiers.

Can I earn money on CVAT?

No — CVAT is an annotation tool used by AI companies, not a contributor pay platform. Data annotators are paid by the companies that use CVAT, not by CVAT itself.

What annotation types does CVAT support?

Object detection, semantic segmentation, image classification, point clouds, 3D cuboids, video annotation, and skeleton labeling — across image, video, and 3D data.

Who uses CVAT?

200,000+ developers worldwide. Notable enterprise users include Diligent, Tecnalia, Intel, EPAM, Vimaan, mBeye, Kitware, and Thermo Fisher Scientific.