Decision Guide

ZenML Open Source vs Cloud

Take advantage of ZenML's core open-source framework available on Github for seamless in-house management, or opt for our feature-rich cloud product designed for effortless collaboration and scalability.
ZenML Open Source vs Cloud

ZenML Cloud is Open Source and More

Built from the foundations the Open Source product, the managed ZenML Pro & Enterprise plans are the best way to scale ZenML usage in your team.

Managed control plane

ZenML Cloud offers multi-tenant, fully-managed ZenML deployments. Seperate your team into workspaces, and deploy dev, staging, and production servers seperately. Your cloud also comes up a managed MLflow playground.

Roles and Permissions

ZenML Cloud tenants have built-in roles and permissions, as an extension to the open-source product. We connect ZenML with your OIDC provider and offer SSO.

Control and configurability

ZenML Cloud control plane allows you to run ZenML pipelines directly from the server, and features enhanced configurability for your pipeline builds.

Enhanced observability

ZenML Cloud tenants have an enhanced dashboard with more features including a model control plane to view all your ML models, and the ability to trigger pipelines, do CI/CD and lots more.
Differences

ZenML Open Source vs Cloud Feature Breakdown

A feature by feature comparison between ZenML Open Source vs ZenML Cloud

(To see differences between ZenML Cloud Pro & Enterprise plans, click here)

Feature

OSS

ZenML Cloud

Pipelines
ML pipelines are Python workflows that execute a machine learning task
Basic Controls with legacy dashboard
Advanced Controls and modern dashboard
Scheduling
Run pipelines on a schedule
Orchestrator-dependant, not available for all orchestrators
Orchestrator-independant managed by ZenML Cloud
Model Control Plane
Machine learning models that are trained or deployed via ZenML
Backend Only
A central view of all your ML models and their associated pipelines, artifacts, and metadata
Artifact Control Plane
Data generated in a ML workflow
Backend Only
A central view of all your data artifacts and their versions
Event Triggers
External sources
Client can trigger the pipeline only
Webhooks to trigger actions (pipeline run, model promote) etc.
CI/CD
Continuous integration and delivery for your pipelines
Self-managed
Integrated with GitHub, Gitlab, Bitbucket. Status checks and event triggers included
Container management
If executed remotely, pipelines run in containers
Basic management
Advanced management with container re-use and optimization
Role Based Access Control
Roles dictate who has permissions to do what
Not available
Fine-grained permissions
User Management
A user account in one ZenML server
Basic
Advanced with SSO
Visualizations & Reports
Visualizing HTML, CSV, Markdown and other sources
Supported via legacy dashboard
Modern dashboard experience
Infrastructure
The infrastructure that supports the central ZenML server
Self-managed
Managed, multi-tenant deployment with database backups, security, compliance, rollbacks, upgrades etc
Service Connectors
Credentials, authorization, and access control for your ML stack components
Legacy dashboard
Modern dashboard
Stacks
Infrastructure components that dictate how a ML pipeline is executed
Legacy dashboard view
Modern dashboard
Integrations
External tools for experiment tracking, model deployment, drift detection, etc.
Community
Purpose-built
Support
Seeking help when stuck
Community
Dedicated 24/7
Setup of MLOps workflow
Setting up of the codebase and infrastructure required to build a successful MLOps platform
Self managed
Specialized onboarding

Start your new ML Project today with ZenML Cloud

Join 1,000s of members already deploying models with ZenML.