Cvat
Open-source data annotation platform for vision AI teams
CVAT is the most flexible open-source choice for vision AI annotation, with options for every deployment need—free self-hosted, managed cloud, or on-prem enterprise. The AI-assisted labeling with SAM and custom models is a real accelerator, and the free tier covers small tests. Watch storage limits on lower plans and note that audio annotation isn't available yet; if you need enterprise support, expect to pay from $12,000/year. For teams migrating from Azure ML Data Labeling before its 2026 retirement, CVAT offers a smooth path.
Verified 3d ago · liveness 73/100 · cite: rightaichoice.com/tools/cvat
- Computer vision engineers building custom models with images, video, or 3D point clouds
- Data labeling teams at autonomous driving, robotics, or manufacturing companies
- Research labs needing an open-source annotation tool for academic datasets
- Enterprises requiring on-premises deployment and data security compliance
- Teams that only need text or natural language annotation tools
- Users who want a fully automated annotation pipeline with no human review stage
- Small teams with extremely tight budgets that can't exceed free-tier storage or project limits
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip CVAT if you need audio annotation today, if you only need text annotation, or if you expect fully automated annotation with no human review stage. Free tier storage and project limits are tight; consider if you can grow beyond them.
On Team plans, SSO, RBAC, and audit logs are only available on annual billing, so you pay yearly upfront to unlock security features.
CVAT's pricing is competitive for teams needing flexibility: Free tier for small tests, Solo at $23–33/mo, Team at $23–33/user/mo, and Enterprise from $12,000/yr. Cheaper than Labelbox or Scale AI for similar capabilities, but Azure ML Data Labeling is being retired, so CVAT is a viable migration target.
In short
Cvat — Open-source data annotation platform for vision AI teams. Best for Computer vision engineers building custom models with images, video, or 3D point clouds, Data labeling teams at autonomous driving, robotics, or manufacturing companies, Research labs needing an open-source annotation tool for academic datasets. Free to start; paid plans from $23/mo.
What's new in Cvat
Checked 3 days agoAcross the latest 1 update: 1 news mention.
What people actually say about Cvat — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
56 mentions across 6 sources (Hacker News, YouTube, Bluesky, Stack Overflow, GitHub, Lemmy) · researched Jul 23, 2026.
- +Open-source with MIT license, no vendor lock-in.
- +Supports image, video, and 3D point cloud annotation.
- +AI-assisted labeling with SAM 2, SAM 3, and Hugging Face.
- +Flexible deployment: self-hosted, managed cloud, or enterprise on-premises.
- +Rich role-based access and project management features.
- −Self-hosted installation is complex and often fails.
- −Nuclio code injection vulnerability (CVE-2025-23045) unaddressed.
- −Built-in QA workflow inadequate, leading to community forks.
- −Managed labeling service has slow response times for applicants.
- −AI polygon feature not consistently available across versions.
- • Managed labeling service has slow turnaround and communication gaps
- • Self-hosted may incur infrastructure and maintenance costs
Viability Score
How well maintained and how widely used is Cvat? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: August 2026
How we score →Key Features
- Image annotation: boxes, polygons, masks, keypoints, skeletons
- Video object tracking across frames
- 3D point cloud detection and tracking (pcd, bin)
- AI-assisted labeling with SAM 2 and SAM 3 for image segmentation
- SAM 2 for video segmentation and tracking
- Integration with Ultralytics, Hugging Face, Roboflow for automatic labeling
- Custom model integration for automatic annotation
- Project, task, and job management with role-based access
- Quality control: ground truth jobs, honeypots, consensus workflows
- Analytics dashboards for progress and workload monitoring
- Import from local files, Amazon S3, Azure Blob Storage, Google Cloud Storage
- Export in 20+ formats (COCO, YOLO, KITTI, Cityscapes, Pascal VOC)
- API, SDK, CLI, and webhooks for automation
- SSO, RBAC, audit logs (Enterprise/Team annual)
- On-premises and air-gapped deployment (Enterprise)
About Cvat
CVAT (Computer Vision Annotation Tool) is an open-source data annotation platform that helps computer vision teams turn raw visual data into model-ready datasets. It supports image, video, and 3D point cloud annotation with tools for detection, segmentation, tracking, and pose estimation. You can import data from local files or cloud storage (Amazon S3, Azure Blob Storage, Google Cloud Storage), organize work into projects, tasks, and jobs with role-based access, and label using manual tools like boxes, polygons, masks, keypoints, skeletons, and tracks. AI-assisted labeling is powered by SAM 2, SAM 3, Ultralytics, Hugging Face, Roboflow, or your own custom models. Quality control features include manual review, ground truth jobs, honeypots, and consensus workflows, with dashboards for progress and workload monitoring. CVAT exports in 20+ formats including COCO, YOLO, KITTI, Cityscapes, and Pascal VOC, and integrates with ML workflows via API, SDK, CLI, and webhooks. Deployment options include self-hosted Community edition (free, MIT license), managed CVAT Online cloud, and CVAT Enterprise for on-premises deployment with SSO, RBAC, audit logs, and air-gapped options. A managed labeling service is also available with 300+ dedicated annotators across 12 time zones, multi-stage QA, and delivery tracking. Audio annotation is listed as coming soon. CVAT is a strong fit for teams that need flexibility, control, and cost predictability, especially those migrating from Azure ML Data Labeling, which retires September 30, 2026.
Behind the Verdict
CVAT stands out in the annotation space for its combination of open-source flexibility, comprehensive tooling, and scalable deployment options. The platform covers the full annotation workflow: import, project management, labeling, quality control, and export—all integrated. The AI-assisted labeling features, including SAM 2 and SAM 3 for image and video segmentation, and integrations with Ultralytics, Hugging Face, and Roboflow, can significantly speed up labeling. Quality control is a strong point, with ground truth jobs, honeypots, and consensus workflows that help ensure high-quality labels. The free tier is ideal for testing, but storage and project limits can quickly become a bottleneck for real projects. Solo and Team plans offer reasonable quotas, but note that SSO, RBAC, and audit logs are only available on annual plans, and Enterprise is required for on-premises deployment. The managed labeling service is a great option for teams that want to outsource annotation entirely. However, audio annotation is still 'coming soon', and the platform is not suited for text or natural language annotation. For teams in autonomous driving, robotics, healthcare, defense, and manufacturing, CVAT is a proven choice. If you are evaluating alternatives, Labelbox and Scale AI offer more automated pipelines but are often more expensive and less flexible in deployment. Overall, CVAT is a robust platform that balances control, cost, and capability better than most commercial tools.
Researching Cvat? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Cvat actually fits — and what changes day-one when you adopt it.
Sign up for CVAT Online Free tier, create a project, upload a small image dataset, and use bounding boxes to label objects for a proof-of-concept model.
Outcome: Start labeling within minutes, export annotations in COCO format, and train a quick YOLO model locally. The free tier's 1GB storage and 3 tasks are sufficient for small datasets.
Subscribe to a Team plan, create an organization, assign roles to annotators and reviewers, and use SAM 2 for video segmentation to accelerate labeling of driving footage.
Outcome: Label and review thousands of video frames with AI assistance, track progress via dashboards, and export in KITTI format. The 30,000 internal AI calls/month cover heavy usage, and webhooks integrate with internal tools.
Deploy CVAT Enterprise on-premises, set up SSO/RBAC, and use air-gapped operation to annotate sensitive medical images.
Outcome: Maintain full control of data, comply with GDPR/HIPAA, and use quality control features like ground truth and honeypots to ensure annotation accuracy. Costs start at $12,000/yr but include custom limits and support.
Use Cases
- Annotate thousands of images for training object detection models in autonomous driving
- Track objects across video frames for surveillance or sports analytics
- Label 3D point clouds for LiDAR-based perception in robotics
- Outsource annotation to CVAT's managed service for large-scale datasets
- Use CVAT Enterprise to annotate sensitive medical imagery on-premises
- Segment anatomical structures in medical scans for diagnostic AI
- Annotate satellite imagery for geospatial mapping and defense
Models Under the Hood
as of 2026-08-21
Limitations
- Free plan supports 1 project and 3 tasks.
- Team plans require a minimum of 2 users.
- Enterprise pricing starts at $12,000 per year.
- Audio annotation is listed as 'coming soon' and not yet available.
as of 2026-08-20
Verification history
We have re-verified Cvat 6 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-checked, vendor evidence unchanged
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Cvat tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Free
$0/mo
Ideal for
Solo evaluators testing CVAT with up to 3 tasks and under 1GB of data; perfect for small proof-of-concepts.
What this tier adds
Free entry point with 1 project, 3 tasks, 1GB storage, 1 cloud storage connection, 100 AI calls/month, and annotations-only export.
Solo Monthly
$33/mo
Ideal for
Individual researchers or freelancers who need more storage (25GB) and project limits, with image export capabilities.
What this tier adds
10 projects, 250 tasks, 25GB storage, 3 cloud storage connections, annotations + images export, and manual review features.
Solo Yearly
$23/mo
Ideal for
Solo users who plan to use CVAT long-term and want to save 30% by paying annually.
What this tier adds
Same features as Solo Monthly but billed yearly at $23/mo, offering cost savings.
Team Monthly
$33/user/mo
Ideal for
Small to medium teams (min 2 users) that need collaboration features like shared projects and more AI calls.
What this tier adds
30 projects, 1250 tasks, 75GB storage, 9 cloud storage connections, 30,000 internal AI calls/month, and webhooks (30). SSO/RBAC only on annual.
Team Yearly
$23/user/mo
Ideal for
Teams committed to using CVAT regularly, wanting lower per-user cost and access to advanced security features.
What this tier adds
Same as Team Monthly but billed yearly at $23/user/mo, including SSO, RBAC, and audit logs (annual only).
Enterprise
Starting at $12,000/year
Ideal for
Organizations with strict security, compliance, or data control needs, requiring on-premises or air-gapped deployment.
What this tier adds
Starting at $12,000/year, includes on-premises/air-gapped deployment, custom limits, SSO/RBAC/audit logs, and dedicated support.
Where the pricing makes sense
The company stage and team size where Cvat's pricing actually pencils out — and where peers do it cheaper.
CVAT's pricing is competitive for teams needing flexibility: Free tier for small tests, Solo at $23–33/mo, Team at $23–33/user/mo, and Enterprise from $12,000/yr. Cheaper than Labelbox or Scale AI for similar capabilities, but Azure ML Data Labeling is being retired, so CVAT is a viable migration target.
Setup time & first value
How long it actually takes to get something useful out of Cvat — broken out by persona, not the marketing-page minute.
For solo users, you can start annotating within 10 minutes of signing up for the free tier. Team admin setup (creating organization, adding users) takes 15-30 minutes. Enterprise on-prem deployment may take 1-2 days, including infrastructure setup and SSO configuration.
Switching to or from Cvat
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From Azure ML Data Labeling: Export labels and data, then use CVAT's import or API to bring them in. Compare formats and use CVAT's 20+ export formats to match your needs.
- ↗To Labelbox: Export CVAT data in COCO or other formats and import into Labelbox. You may need to map label schemas.
- ↗To Scale AI: Use CVAT's export formats to transfer datasets, then manually set up projects in Scale.
Integrations
Resources & Guides
Tutorials & Learning
Official links
Tools that pair well with Cvat
Common stack mates teams adopt alongside Cvat, with the specific reason each pairing earns its keep.
Featured Head-to-Head Comparisons
Cvat vs Presto Voice
Presto Voice and CVAT serve entirely different purposes: Presto drives revenue for QSR chains through voice AI order automation and upselling, while CVAT is an open-source annotation tool for vision AI teams. Choose Presto if you run a multi-location drive-thru chain and want proven ROI; choose CVAT if you need flexible, self-hosted data annotation for computer vision projects.
Cvat vs Screenplayiq
ScreenplayIQ is a niche tool for film industry professionals needing data-driven script financial forecasts, while CVAT is a versatile open-source annotation platform for computer vision teams. Choose ScreenplayIQ if you're a screenwriter or studio exec evaluating feature film scripts; choose CVAT if you're building machine learning models that require labeled visual data. They serve completely different domains, so the decision hinges entirely on your need for screenplay analysis versus visual data annotation.
Cvat vs Truleo
Truleo and CVAT serve entirely different needs. For law enforcement agencies drowning in siloed data, Truleo is a specialized intelligence platform that cuts report writing time from 40 to 7 minutes. For computer vision teams building custom models, CVAT is a flexible, open-source annotation tool with AI-assisted features. Choosing between them depends solely on your domain: policing or vision AI.
Alternatives to Cvat
View allFrequently Asked Questions
Used Cvat? Help shape our editorial sentiment research.