Open-source pipeline framework that integrates all your ML tools
Get all your ML workflows running on any tooling stack with minimum effort.
Author Your Pipeline
ZenML provides an easy-to-use interface to create tool and infrastructure agnostic pipelines for batch processing, continuous training, and deployment. Quickly build, execute and interact with completed runs through ZenML.
HOW IT WORKS
Get started quickly
ZenML is completely free and open-source. See the magic with just two simple commands:
$pip install zenml
$zenml go
PROBLEM SOLVING
ZenML Features
Bring your own tools agnostic of your pipeline
ZenML lets you integrate with your favourite tools. You can alsoimplement your own integration.
Deploy MLOps Stacks With ZenML
Quickly deploy popular MLOps stacks for your ZenML pipelines by simply using:
$pip install zenml
$zenml stack recipe deploy aws-minimal
Our Roadmap
Why our users love ZenML
Chris Manning
@chrmanning
Many, many teams still struggle with managing models, datasets, code, and monitoring as they deploy ML models into production. ZenML provides a solid toolkit for making that easy in the Python ML world
Professor of Linguistics and CS at Stanford
Goku Mohandas
@GokuMohandas
ZenML allows you to quickly and responsibly go from POC to production ML systems while enabling reproducibility, flexibitiliy, and above all, sanity.
Founder of MadeWithML
Richard Socher
@RichardSocher
ZenML allows orchestrating ML pipelines independent of any infrastructure or tooling choices. ML teams can free their minds of tooling FOMO from the fast-moving MLOps space, with the simple and extensible ZenML interface. No more vendor lock-in, or massive switching costs!
Chief Scientist Salesforce and Founder of You.com
Matt Squire
@FuzzyLabsAI
A lot of our teams struggle to bring sanity to their model training processes. ZenML is built in a way that encourages good, maintainable pipelines. It makes it easy to coordinate all the various tools, systems and assets that go into training ML models and, crucially, it gives you as much control over the process as you need.
CTO at Fuzzy Labs
Gabriel Martin
@gabrielmbmb_
ZenML allows you to keep your ML pipeline code cloud-agnostic, enabling faster future migrations to another technology stack. The management of the metadata and artifacts generated at each step is seamless, and allows the user to extend the framework if needed without much effort.
Machine Learning Engineer at Frontiers
Frequently Asked Questions
Canโt find the answer to your question?
Q: Why should I use ZenML?
The MLOps landscape is complex and hard to navigate. If the wrong decisions are made while creating your technology stack early on, it leads to complexity and technical debt as you progress through the stages of MLOps maturity. ZenML allows you to define your ML workflows in a vendor- and tool- agnostic manner, so that you can write your code knowing that you can switch out ML components easily as and when it is needed. Read more our our blog.
Still not clear?
Send us a message
Q: How does ZenML compare to orchestrators (Airflow, Kubeflow, Flyte, etc.)?
At a high level, an orchestrator in MLOps is a tool that enables developers to schedule, monitor, and manage workflows. ZenML is not an orchestrator, but rather is a tool that lets you author pipelines that can be run on multiple orchestration systems. There are standard orchestrators that ZenML supports out-of-the-box, but you are encouraged to write your own orchestrator in order to gain more control as to exactly how your pipelines are executed. ZenML not only allows you to easily swap our orchestrators, but also then focuses on adding value within the pipeline itself, with features such as model deployment, metadata tracking, and management of the configuration of your MLOps stack.
Still not clear?
Send us a message
How can I quickly deploy the infrastructure required to run ZenML?
There are two sides to this question. ZenML itself can be deployed as a service, which centrally tracks everything related to pipelines and stacks. However, ZenML also often requires additional infrastructure for various stacks. The ZenML MLOps Stack repository contains Terraform scripts to quickly get you started with these stacks. These recipes can also be called directly with ZenML with zenml stack recipe deploy . Read more here.
Still not clear?
Send us a message
Q: How can I be sure you'll stick around as a tool?
The team behind ZenML have a shared vision of making MLOps simple and accessible to accelerate problem solving in the world. We recently raised our seed round to fulfill this vision, and you can be sure we're here to stay! Plus, ZenML is and always will be an open-source effort, which lowers the risk of it just going away any time soon.
Still not clear?
Send us a message
Q: Can I use tool X? How does tool Y integrate with ZenML?
Take a look at our examples directory on the GitHub, which showcases detailed examples for each integration that ZenML supports out-of-the-box. The ZenML team and community is constantly working to include more tools and integrations to the above list (check out the roadmap for more details). You can upvote features you'd like and add your ideas to the roadmap (https://zenml.io/roadmap)
Still not clear?
Send us a message
Canโt find the answer to your question?
Read more on our Docs
ZenML MLOps stack discovery
Find out if ZenML supports your MLOps stack
Fill out this short questionnaire to tell us about the tools you already use, and we'll get back to you with a concrete suggestion on how you can start benefiting from ZenML!
ZenML CLOUD
Manage MLOps for teams with ZenML Cloud
Centralized tracking, pipeline registry and managed backend for ZenML
Join our newsletter
Sign up to our โBuilding an open source companyโ newsletter by the ZenML team.
Weโre building tools for production machine learning, and planning to share the startup journey as we walk through it.
