CVAT
An open-source platform for annotating images, videos, and 3D data to train machine learning models.
| What is it | An open-source platform for annotating images, videos, and 3D data to train machine learning models. |
|---|---|
| Pricing | Freemium — from $33/mo |
| Free tier | Yes |
| Platform | Web Application |
| API | Yes |
| Best for | labeling medical images for disease detection models, annotating video for autonomous vehicle training data |
| Domain registered | 2022 |
Data updated Aug. 1, 2026
What does CVAT do?
CVAT is a data annotation platform built specifically for creating training data for machine learning and computer vision models. It provides a web-based interface where teams can manually or automatically label images, videos, and 3D point cloud data. The core task is drawing bounding boxes, polygons, points, and other markers on visual data to teach AI models what to recognize.
It works by letting users upload datasets, often from cloud storage like AWS S3, and use a variety of annotation tools. A standout feature is auto-annotation, where you can use a pre-trained AI model or your own to suggest labels, speeding up the process. CVAT supports a wide array of annotation types, from simple image classification to complex 3D cuboids and video object tracking. It also exports data in 19 different formats, making it compatible with most ML frameworks.
This tool is most useful for professional teams building AI in fields like healthcare, retail, automotive, and manufacturing. For example, a medical research team can use it to label X-rays to train a model for early disease detection. A company developing self-driving cars can annotate hours of video footage to teach its system to identify pedestrians and traffic signs. It's built for scale, handling large datasets and providing management analytics to track annotation progress and team productivity.
Key features
What makes it stand outWho is CVAT for?
Who benefits most from this toolPricing
Free tier available — start without a credit cardFree
- 3 tasks
- 1 projects
- 1 cloud storage
- 1 GB internal storage
- 100 external AI calls
- 100 internal AI calls
- Image support
- Video support
- 3D support
- API Access
- Data export (annotations only)
- Webhooks (1)
- Internal AI calls (100/month)
- External AI calls (100/month)
Solo
- 250 tasks
- 10 projects
- 3 cloud storage
- 25 GB internal storage
- Fair usage external AI calls
- 10,000 internal AI calls
- Image support
- Video support
- 3D support
- API Access
- Data export (annotations + images)
- Webhooks (10)
- Internal AI calls (10,000/month)
- Projects (10)
- Tasks (250)
- Internal storage (25 GB)
- Cloud storage (3)
Team
- 23 price per user yearly
- 33 price per user monthly
- Team collaboration
- Image support
- Video support
- 3D support
- API Access
- Data export (annotations + images)
- Manual review
- Ground truth job
- Honey pot
- Reports & analytics
- Role-based access controls
- Audit logs
- Webhooks
- Internal AI calls
- External AI calls
- Automatic annotation
- Integration with Hugging Face & Roboflow
- SAM 2 and SAM 3
- SAM 2 for video
CVAT for Enterprises
- 12,000 minimum yearly price
- Host CVAT securely within your own infrastructure
- Enhanced security
- Compliance
- Control over data
- Suitable for teams of all sizes
Trust & presence
Gallery
Click any image to enlargeAlternatives in Image & Art
AI-powered browser extension and web app to collect, organize, and search your image library with one click.
Convert 2D images into immersive 3D experiences viewable on any device without special glasses
AI-powered digital asset manager for teams — organize, edit, and host photos in one collaborative workspace.
Free AI font generator — type your text, choose from 100+ styles, and download a PNG/JPG instantly.
Platform to build, train, and deploy computer vision models — manage datasets, fine-tune models, and go to production.
Free gallery of pre-made AI image prompts — copy and use instantly with GPT Image 2, Nano Banana, Midjourney, and other models.
A curated gallery of creepy and unnerving AI-generated images for Halloween inspiration and spooky content.
A free weekly newsletter with the latest generative AI art news, artist interviews, and resources.
Similar tools
Data annotation services for AI teams — image, video, 3D, and text labeling to train computer vision and machine learning models.
Automated data annotation platform for computer vision — uses AI to label images 15x faster with precision.
A web-based platform for labeling images, videos, and 3D data to train machine learning models.
All-in-one smart data annotation platform for creating training datasets for AI models across multiple data types
AI-powered platform for accelerating computer vision model development with automated data labeling and quality assurance.
Multi-sensor data labeling platform for robotics and autonomous vehicles — label 2D images and 3D point clouds together.
AI-powered data annotation tool — upload images, mark objects once, get batch annotations instantly
Open source platform for labeling data to train, fine-tune, and evaluate AI models across text, images, audio, and video.