kitio.aiAdded Nov 5, 2026
Kubeflow
FreeOpen-source Kubernetes stack for training and serving AI models
What is Kubeflow?
Kubeflow is a set of open-source projects that run machine learning and AI workloads on Kubernetes, from notebooks and training to pipelines and model management. Platform teams use it to give data scientists the same tooling on a laptop, on-premises cluster or any cloud.
Subprojects include Kubeflow Notebooks for interactive environments, Trainer for distributed training and LLM fine-tuning across PyTorch, JAX, DeepSpeed and other frameworks, Katib for hyperparameter tuning and AutoML, Pipelines for portable workflows, and Hub for tracking model versions and artifacts.
A Spark Operator runs Spark jobs on the same clusters, and a central dashboard ties the web interfaces together. Kubeflow is a graduated Cloud Native Computing Foundation project, developed in the open with weekly community calls, a mailing list and Slack.
Features
- Kubeflow Pipelines
- Distributed training
- Hyperparameter tuning
- Hosted notebooks
- Model registry
- Spark Operator
Tags
Releases
+ Submit new release- Listed
Kubeflow was added to the directory.
Organization
Kubeflow community (CNCF)
Listed by kitio.ai · unverified owner
Pricing
- Pricing model
- Free
- Paid options from
- 100% free
Sponsored
Your AI tool here
A Featured launch puts your tool in this spot and on the home page for a week.
Top Kubeflow alternatives
View allModels & APIs
Open-source control plane for ML training, serving and agents on Kubernetes
Models & APIs
Open-source Python framework for ML and AI workflows
Models & APIs
Open-source platform for ML experiments, models and GenAI tracing
Models & APIs
Open-source MLOps and AI infrastructure platform for GPU clusters
If you liked Kubeflow
- NannyMLNew
Monitors deployed ML models and estimates performance without labels
FreemiumModels & APIsMade in Belgium - MLServerNew
Open-source Python server for serving ML models over REST and gRPC
FreeModels & APIsMade in United Kingdom Open-source model serving and a managed platform for AI inference
FreemiumModels & APIsMade in United States
Pros and cons
Pros
- Free and open source
- Runs on any Kubernetes cluster
- Modular, use only the parts you need
Cons
- Requires Kubernetes knowledge to operate
- Setup and upgrades take real effort
- No official managed version from the project
Prompts & results
No prompts yet. Add your own prompt and output to help others use this tool.
Share a prompt
Reviews
Q&A
What is Kubeflow?
Kubeflow is a set of open-source projects that run machine learning and AI workloads on Kubernetes, from notebooks and training to pipelines and model management. Platform teams use it to give data scientists the same tooling on a laptop, on-premises cluster or any cloud.
Can I use only part of Kubeflow?
Yes. Kubeflow is modular, so you can deploy individual subprojects such as Pipelines or Trainer without the full stack.
Which training frameworks does Kubeflow Trainer support?
It works with PyTorch, MLX, Hugging Face, DeepSpeed, Megatron, JAX and XGBoost, among others.
How much does Kubeflow cost?
Kubeflow is 100% free.
Is there a free version of Kubeflow?
Yes. Kubeflow is completely free to use.
What are the main features of Kubeflow?
Kubeflow Pipelines, Distributed training, Hyperparameter tuning, Hosted notebooks, Model registry, Spark Operator.
Who makes Kubeflow?
Kubeflow is made by Kubeflow community (CNCF).
What are the best alternatives to Kubeflow?
Popular alternatives include Polyaxon, Metaflow, MLflow.