Get the best MLOps experience

for teams of any size

FREE

Best for individuals, researchers, academia, and small teams
(up to 3 users)

$0

No credit card required

Experiment Management

Model Repository

Artifacts

Dataset Versioning​

Pipelines

Agent Orchestration

CI/CD Automation

Reports

100GB Free artifact storage
1GB Metric Events
1M API calls

* Metric Events are console logs, scalars & plots. 1GB of events equals to the total amount of storage required to save all metric events.

START FOR FREE

PRO

Enhanced features for
growing teams
(up to 10 users)

$15

Per User/Month + Usage

Everything from Free​ + Applications:

Cloud Auto Scaling

Hyperparameter Optimization

Task Scheduling​

Pipeline Triggers​

Dashboards​

Alerts

+ 20% included free
$0.1 / 1GB artifact storage
$0.01/1MB Metric Events
$1/100K API calls
$0.04/hr per Application

Get Started

SCALE

Best for scaling ML pipelines and data abstraction, with increased control & support

Custom

Everything from Pro

VPC - Private Server

HyperDatasets i

Kubernetes Integration

SSO

Support

SLA

Unlimited usage
Request a Demo

ENTERPRISE

Unmatched performance, scalability and access control for production ML

Custom

Everything from Scale

VPC / On-PREM - cluster

ClearML Custom Apps​

LDAP integration​

Role-based access control​

Whiteglove Support​

Professional Services​

Custom SLA

Unlimited usage
Request a Demo

You can also run self-hosted ClearML

100% Open Source

What plan should I choose?

Features

Open Source

Free / Pro

Scale

Enterprise

Experiment Tracking

2-line orchestration

Convert any code/repo into fully orchestrated DevOps solution with only 2 python lines

Comparison

Artifacts

Search capabilities

Including programmatic interface

Data Samples

Store any media / file for better visibility and debugging capabilities

Scalars

Plots (2D/3D/Custom)

Project Dashboard

Git Version Control integration

Includes storing any uncommitted changes

Package Tracking

Automatically track any python packages used including specific version

Log Command line & Console

Track and log command line arguments and console output

Configuration Files

Automatically connect with configuration packages (hydra) or any custom configuration file

Share Direct Links

Easily collaborate with colleagues, share address url to reproduce the exact UI view on any machine

Projects & Sub-Projects​

Easily collaborate with colleagues, share address url to reproduce the exact UI view on any machine

Permissions

Role Based permission access control, for projects & experiments

Reports

Create and share reports within the platform and embed real-time graphs in third-party editors

Agent Orchestration

Spin experiments in development & production within any container with minimum effort using bare-metal or Kubernetes

Workers

Unlimited

2 / Unlimited

Unlimited

Unlimited

Pipelines

Programmatic pipelines grant easy integration and debugging with full customization support

CI / CD automation

Integrate CI/CD actions with an easy-to-use RestAPI and Python interface, while also supporting continuous training action based on model performance

Job Scheduling

Schedule jobs with priority queues and resources, and full integration with Web UI

Dynamic Job Scheduling

Dynamic bare-metal job scheduling including dynamic GPU allocation and opportunistic allocation

Configuration Vault

Store configuration vault per user / group / company, supporting environment variables, access keys, storage credentials etc.

Optimal container selection

Customize container matching based on required environment, packages, python, projects, etc.

Scheduling Quota & Policy Manager

Dynamic graph based decision tree for job allocation. Per user/group quota. Optimal execution queue matching, base container, projects, groups, etc. Dynamic resource optimization, online job scheduling between different resources (local/cloud/k8s)

Dataset versioning

Fully differentiable data store with transparent version control on top of any storage

Dataset management

Add meta-data (tags, names, descriptions, projects) on any dataset and version , to create hierarchical managed data

Dataset Searchability

Search datasets / versions based on names / comments / tags etc

Programmatic Data Previews

Add sample data representation viewable in the Web UI, for increased data visibility

Metadata versioning

Immutable versions

Dataset version graph

Multiple version inheritance

Combine dataset version from multiple parents

Differentiable storage

Versions only store the change set from parent versions

Flexible storage

Store your datasets on any storage and connect to ClearML – Http/S3/GS/Azure/NAS

Local Cache support

Hyper-Datasets Management

Data access abstraction layer fully separating code from data, along with DB query capabilities and version control built-in

File level Queryable MetaData

DB query capabilities on any file level meta-data

Data / Meta visualization

Full Data preview inside the UI, including metadata

Image/Video/Time-series/Text annotations visualization

Review visual annotation on top of images/videos

File level meta-data & annotation Web UI edit

Modify fix and add file level annotation

Combined queries from multiple dataset versions

Online dataset rebalancing

Oversample data based on any query, rebalancing dataset in real-time based on metadata

Individual file links

Link pre-uploaded files, supports S3/FS/Azure/Nas/Http

Stored queries (DataView)

Store dataset query as individual entities

Experiment Management integration

Automatically log and modify dataset on an experient level

Model repository

Automatically log and upload any model in the system into a unified searchable model repository that is ready for production

Query capabilities

Search and filter models based on projects/names/tags/meta-data

Model management: state/ tag /project / name / description

Add searchable meta-data on any model version (tag/name/etc.), to create searchable hierarchy managed models

Automatic DL/ML Framework integration

Automatic integration with leading DL/ML frameworks (upload/download model files)

Model meta-data key-val/comments

Model genealogy

Connect the creating process with the model. Log previous iteration of a model (coming soon)

Trigger action

Trigger external action based on model repository changes (publish/tag/ etc.)

Basic

Serving

Spin and manage model serving engines with automatic integration to the ClearML model repository (beta)

HyperParameter Optimization

State-of-the-art bayesian optimization algorithms help find the best hyper parameter combination running on multiple machines with an easy-to-use web UI

w/o UI

Pro Only

Task Scheduling

Schedule any pythonic code to be repeated in predefined intervals. Set base container, required packages and have full monitoring & resource control, directly from the web UI

w/o UI

Pro Only

Pipeline Triggers

Triggered automation pipelines (or any pythonic code) based on changes in the system, new dataset created, model published, task reporting specific metrics, etc. Think Git actions tailored towards ML/DL workflows

w/o UI

Pro Only

Project Dashboard

Create custom dashboards for your project, monitor progress, summarize metrics over time, review failed experiments and create alerts directly connected with your internal Slack channels

w/o UI

Pro Only

Task Monitoring

Shutdown active experiments or tasks based on resource monitoring metrics or performance metrics with full integration with Slack for shutdown alerts (beta)

Pro Only

ClearML Deploy

Model deployment architecture leveraging sota model serving engine, allowing for customized endpoints, node autoscaling, and full monitoring (beta)

w/o UI

Coming soon

ClearML Remote Development

Allocate remote machines (GPU/CPU) for development with a single CLI command, supporting both bare-metal and Kubernetes

Pro Only

ClearML Custom Application

Build your own application on top of ClearML. Create easy to utilize data science workflows, expanding BU abilities to use your custom models and pipelines.
Fully customizable launch wizard with programmatic interface and custom real time dashboard.

Flexible Deployment

Deploy ClearML-server on any infrastructure and attach compute nodes (CPU/GPU) from multiple infrastructures: bare-metal, Kubernetes, and leading cloud vendors.
BYO

SaaS

VPC- Single Server

VPC / On-Prem Cluster

Kubernetes Integration

Integrate resource scheduling into your existing Kubernetes cluster(s). Support combination with additional resources (bare-metal/cloud)

Basic

Rest API

Security

BYO

Role-based Access Control

Single Sign On

Google/GitHub/BitBucket/Okta/Auth0, etc.

LDAP Integration

Dedicated Support

Customization, DevOps, MLOps, and data science support

SLA

Don't see the right plan?

Get in touch with our team and we will assist you with building your business’s custom ClearML license.

Let's Chat

Powered by Open Source. Free Forever. Made with ❤ by ClearML

Product

Product

Company

Resources

Follow

Github Slack YouTube Facebook Twitter LinkedIn
@2023 • Allegro AI. All Rights Reserved

Terms of Use    Privacy Policy