Supervisely
supervisely.com
- Free plan
- Web
- API
- Self-hosted
What Supervisely is
Supervisely is a computer vision platform for managing datasets, labeling images and other media, and developing production models. Its annotation workflows cover images, video, LiDAR, and DICOM, with AI-assisted labeling and tracking. Teams can train, evaluate, and deploy models, including through active-learning workflows. Data tools provide search, visualization, and quality statistics, with connections to AWS, Google Cloud, and Azure storage. A JSON API, Python SDK, and AppEngine support workflow automation and integrations. The Community plan costs nothing and includes 5 GB of storage, 10,000 files, and two members; projects in free teams are archived after 30 days of inactivity. Pro is €199 a month, with 50 GB storage and 50,000 files; additional capacity is available for a fee. Enterprise has custom pricing and can be self-hosted or cloud-deployed, including on private networks without outside internet access. The pricing page says customer data is private and is not shared or used by Supervisely.
Who it suits
Community is described for open-source projects, individuals, ML researchers, and small teams. Enterprise deployment and support options suit organizations with private-network requirements or custom needs.
Good to know, and what to watch
- Covers image, video, LiDAR, and DICOM labeling.
- Includes AI-assisted labeling and tracking.
- Supports training, evaluation, and model deployment.
- Offers API, Python SDK, and AppEngine.
- Enterprise can run on private networks without internet.
- Community has 5 GB of storage and two members.
- Free-team projects archive after 30 days inactive.
- Pro capacity add-ons cost extra.
Plans
Every plan and priceEnterprise
Community
Pro
Free plan · 30-day trial
Getting started
- Open the Supervisely website
- Choose the web, API, or self-hosted route
- Start with the Community plan, the Pro plan, or Enterprise
- Use the 30-day Pro trial if evaluating that plan
- Connect AWS, Google Cloud, or Azure storage when needed
- Use the JSON API, Python SDK, or AppEngine for integrations
Limits to check
Community includes 5 GB storage, 10,000 files, and two members, with projects archived after 30 days of inactivity. Pro includes 50 GB and 50,000 files; extra storage and file capacity are available for a fee.
The Club's review
Supervisely is a computer-vision workspace for teams that need to move from raw visual data to trained and deployed models. It suits ML researchers, open-source contributors, and growing engineering teams with varied annotation needs; organizations requiring private-network deployment have an Enterprise path. The trade-off is clear: the free tier is small, while serious shared production work starts at €199 per month.
Overview
Supervisely combines dataset curation, labeling, and model operations rather than treating annotation as an isolated task. You can work with images, video, LiDAR, and medical DICOM imagery, then train, evaluate, and deploy models in the same environment. Active-learning workflows are included for projects that want model output to guide the next labeling cycle.
Search, visualization, and quality statistics make the data layer useful before and after annotation. Connections to AWS, Google Cloud, and Azure storage help teams keep cloud-based datasets in their existing estates. Supervisely dates its product focus and first release to 2017, and the breadth is aimed at computer-vision programs rather than occasional one-off labeling.
Key features
Annotation and formats
The labeling toolkit covers rectangles, lines and polylines, polygons, multipolygons, points, bitmap masks, alpha masks, graphs and keypoints, and 2D cuboids. AI-assisted labeling and tracking can reduce repetitive work, especially in image and video projects. Support for common image types—including JPEG, PNG, WebP, TIFF, AVIF, and HEIC—handles typical camera and document pipelines.
Exports include Supervisely Format, COCO, Pascal VOC, YOLOv8, DOTA, and Cityscapes, plus direct destinations such as Amazon S3, Google Cloud Storage, Microsoft Azure, and S3-compatible storage. That range is practical when downstream training or labeling tools impose a particular dataset convention.
Models, automation, and collaboration
Training, evaluation, deployment, and active learning give teams a continuous model workflow instead of a separate annotation handoff. A JSON API, Python SDK, and AppEngine support scripted operations and custom integrations. Team collaboration is supported, with role-based permissions for controlling access.
Security and deployment
Enterprise can run as self-hosted software or as a cloud deployment, including on private networks without outside internet access. Security controls include role-based access, SSL certificates, and optional threat scanning for uploaded assets. Customer data is private and is not shared with or used by Supervisely. Enterprise support adds priority assistance, a personal success manager, and one-to-one training calls.
Pricing
| Plan | Price | What you get | Best for |
|---|---|---|---|
| Community | Free | 5 GB storage, 10,000 files, two members, unlimited projects and annotations | Individuals, open-source work, ML researchers, and small teams |
| Pro | €199 a month | 50 GB storage, 50,000 files, 30-day free trial, paid add-ons | Teams needing more capacity and shared production work |
| Enterprise | Custom pricing | Unlimited storage and files, self-hosted or cloud deployment, custom requirements | Organizations with scale, compliance, or private-network needs |
Community is genuinely usable for small projects, but two seats and 5 GB constrain shared datasets quickly. Free-team projects are archived after 30 days of inactivity, so it is a poor fit for dormant archives that must remain immediately active. Pro raises the storage and file ceilings substantially; extra capacity costs more, and the €199 monthly entry price makes it a considered purchase rather than a casual upgrade. Enterprise is the route when infrastructure control, unlimited capacity, or tailored requirements matter.
Platforms
Supervisely is offered through the web, an API, and self-hosted deployment. The API and Python SDK suit automated pipelines, while self-hosting is specifically available for Enterprise environments that cannot permit external internet access.
Who it's for
Choose Supervisely when your computer-vision program spans multiple annotation types, cloud data stores, and model development. It is particularly well matched to research groups and small teams starting on Community, then moving to Pro as storage, files, and collaboration grow. Enterprise is the sensible choice for companies that need private deployment and formal support. If you only need a lightweight, single-purpose labeler, the platform's broader workflow and higher paid entry point may be unnecessary.
Pros and cons
Pros
- Broad visual coverage: image, video, LiDAR, and DICOM workflows handle more than standard 2D datasets.
- End-to-end workflow: curation, annotation, active learning, training, evaluation, and deployment are connected.
- Strong interchange: major computer-vision export formats and three large cloud-storage providers support existing pipelines.
- Enterprise control: self-hosting, private-network operation, permissions, and security options address sensitive deployments.
Cons
- Small free footprint: 5 GB, 10,000 files, and two members limit Community for larger teams.
- Inactive-project archiving: free-team projects archive after 30 days without activity.
- High Pro starting cost: €199 per month is difficult to justify for occasional or hobby use.
- Capacity costs extra: Pro users pay for additional storage and file capacity.
Alternatives
Label Studio is a freemium option with Community Edition and broad API, desktop, web, and self-hosted platform coverage; choose it when platform variety is the priority.
Segments.ai offers a free academic license for non-commercial research, making it a better fit for eligible students and researchers.
Roboflow has a free tier with monthly credits and supports mobile platforms, so it suits teams preferring credit-based usage and deployment reach.
CVAT provides a free MIT-licensed community edition and a limited online free plan, useful for personal use or small teams seeking an open option.
MakeSense.ai is a free browser-based photo-labeling tool for straightforward labeling without a broader model platform.
Labellerr offers a free researcher plan with one seat and limited credits, which may suit small academic projects.
Draw Attention targets interactive images rather than a full computer-vision development workflow.
Ultralytics Platform has a free plan with substantial stated storage and model-training allowances, worth considering when those quotas match your workload.
For broader comparisons, see Image Annotation Software, AI Data Labeling Tools, AI Image Annotation Tools, Data Labeling Software, and Video Annotation Software.
Verdict
Supervisely is a strong choice for teams that want one environment for diverse visual annotation, governed datasets, and model workflows. Its API, cloud connections, and Enterprise private deployment add practical scale. Look elsewhere if your needs stop at simple labeling or if €199 per month and the Community limits do not fit your budget or team size.
The verdict
Supervisely brings data curation, annotation, and model workflows into one computer vision platform. Its free plan has defined storage and team limits, while Enterprise supports self-hosted deployment and custom requirements.
Questions people ask
Does Supervisely have a free plan?
Yes. Community costs nothing and includes 5 GB storage, 10,000 files, and two members.
How much does Pro cost?
Pro is listed at €199 a month and includes a 30-day free trial. Add-ons are available.
Can Supervisely be self-hosted?
Yes. Enterprise Edition is available as self-hosted or cloud deployment and can run on private networks without outside internet access.
What integrations does Supervisely provide?
It offers a JSON API, Python SDK, and AppEngine, plus connections to AWS, Google Cloud, and Azure storage.
Who is the Community plan for?
The Community plan is described for open-source projects, individuals, ML researchers, and small teams.
Is customer data private?
The pricing page says uploaded or generated customer data is private and is not shared or used by Supervisely.
What the maker publishes
- Annotation
- Its labeling tools include image, video, LiDAR and DICOM workflows, with AI-assisted labeling and tracking. source
- Company history
- Supervisely says it shifted its primary focus from services to product in 2017 and released Supervisely in August 2017. source
- Data management
- It supports data search, visualization, quality statistics and connections to AWS, Google Cloud and Azure storage. source
- Data privacy
- The pricing page says uploaded or generated data is private to the customer and is not shared or used by Supervisely. source
- Enterprise deployment
- Enterprise Edition is available as a self-hosted or cloud deployment and can run on private networks without outside internet access. source
- Enterprise support
- Enterprise support includes priority support, a personal success manager and one-on-one training calls. source
- Founded
- 2017 source
- Free plan limits
- Community includes 5 GB of storage, 10,000 files and two members, and projects in free teams are archived after 30 days of inactivity. source
- Headquarters
- Tallinn, Estonia source
- Integrations
- Supervisely offers a JSON API, Python SDK and AppEngine for automating workflows and building integrations. source
- Intended users
- The Community plan is described for open-source projects, individuals, ML researchers and small teams. source
- Model workflows
- The platform includes tools to train, evaluate and deploy models, including active-learning workflows. source
- Product
- Supervisely is a computer vision platform for curating data, labeling images, videos, 3D and medical imagery, and building production models. source
- Pro limits
- The pricing page lists Pro at 50 GB storage and 50,000 files, with additional storage and file capacity available for a fee. source
- Security
- The security page describes role-based access permissions, SSL certificates and optional verification of uploaded assets with threat scanners. source
Checked against the maker's pages on 4 Oct 2026.