Kubeflow
A cloud-native platform for data, AI/ML, and HPC workloads.
- Source repo
- kubeflow/kubeflow
- Stars
- ★ 16k
- Last updated
- 1mo ago
- License
- Apache-2.0
- FA score
- 28/100 · Major gaps
At a glance
- How it runs
- Works with
- Universal · cross-platform
- Cost
- Free, no paid service needed
- Setup effort
- High · needs real infrastructure
- You'll need
- Typical use
- AI practitioners running scalable AI/ML workloads on Kubernetes infrastructure.
- Not a fit if
- Teams seeking one monolithic application instead of a modular project stack
- Users expecting complete installation commands in this repository
- Teams that do not plan to adopt Kubernetes
- Website
- kubeflow.org ↗ set on GitHub
- Maintenance
- Commits in the last 6 monthslast commit 2026-09-30latest release redirect (2026-04-29)+1.4 ★/day over 7 days
- Source review
- 28/100 · Major gaps 7 safety controls not found
What does this agent do, and when should you use it?
Kubeflow is a Kubernetes-native platform assembled from modular open-source subprojects for data and AI workloads. It serves AI practitioners, platform administrators, and decision-makers, with stated scope spanning data, AI/ML, HPC, large-scale training jobs, and high-throughput AI agents. Its composable design lets teams select and combine tools across the AI lifecycle. The same code is intended to run on a local laptop, on premises, or in a cloud, with Kubernetes defining the deployment boundary. This repository is primarily a gateway to the subprojects and shared project metadata; implementation work happens in separate subproject repositories, so it is not a complete installation or product implementation by itself.
Kubeflow organizes modular open-source projects into a Kubernetes-native stack for data and AI workloads. The platform is intended to run workloads ranging from data, AI/ML, and HPC use cases to hyperscale training jobs and high-throughput AI agents; individual subprojects supply the concrete execution components and interfaces. Its composable model allows teams to mix tools across the AI lifecycle and run the same code on a laptop, on-premises infrastructure, or cloud infrastructure. This gateway repository does not expose a verifiable unified API, CLI command, input contract, or output format in the supplied material.
- AI practitioners running scalable AI/ML workloads on Kubernetes infrastructure.
- Platform administrators building a modular shared environment for data, machine-learning, and HPC teams.
- Training teams managing hyperscale jobs while retaining a path across local, on-premises, and cloud environments.
- Agent-platform teams needing Kubernetes-native infrastructure for high-throughput AI-agent workloads.
- Technology decision-makers assembling open-source tools for different stages of the AI lifecycle.
How do you install or deploy this agent?
The supplied material provides no copyable installation command, supported Kubernetes version, cluster-preparation procedure, distribution-selection process, or first-run example. Because this repository is a gateway to Kubeflow subprojects and shared metadata, an accurate end-to-end installation cannot be derived from it. Deployment requires selecting the relevant subprojects or a Kubeflow distribution and following installation material not included in the source.
How do you use this agent?
The source does not document a unified CLI, API, configuration file, credential requirement, or first working invocation. The evidenced usage model is to combine Kubeflow subprojects as needed across the AI lifecycle and run their workloads on a Kubernetes-native stack deployed on a laptop, on premises, or in a cloud. Procedures for submitting a training job, opening a notebook, running an agent, or retrieving output are not provided.
What are this agent's strengths and limitations?
- Its modular architecture lets adopters combine open-source subprojects for particular stages of the AI lifecycle.
- It explicitly targets portability across local laptops, on-premises environments, and clouds.
- Kubernetes provides a common execution foundation for scaling to hyperscale training jobs and high-throughput AI agents.
- Its stated scope includes data, AI/ML, and HPC workloads rather than a single model-development stage.
- The Kubernetes-native architecture requires adopters to provide and operate real cluster infrastructure.
- This repository is mainly a gateway, while implementation is distributed across separate subproject repositories, increasing evaluation and maintenance scope.
- The supplied material lacks installation commands, version requirements, configuration steps, and a runnable first example.
- No unified API, CLI, failure-handling process, or compatibility boundaries between subprojects are documented here.
How does this agent compare with similar options?
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| Kubeflow This agent | 28 · Major gaps | Self-hosted serviceFree | ★ 16k | 1mo ago | — | — |
| AgileRL | 62 · Some gaps | Library / SDKFreemium | ★ 970 | today | Python | — |
| Metaflow | 60 · Some gaps | Library / SDKFree | ★ 10k | 4d ago | Python | — |
| ML Road: Machine Learning & Agentic AI Resource Collection | 29 · Major gaps | Web appFree | ★ 5k | 3mo ago | Python | — |
How does FollowAgents rate this agent?
Why each dimension lost points
The supplied files do not show Agent permissions, confirmation gates, data flows, sensitive data handling, dependency security, external side effects, or rollback, so those criteria score 0. The README identifies Kubeflow as a Kubernetes-native AI platform and describes community maintenance, but provides no specific maintainer or verified source provenance; attribution is only partly traceable and scores 1.
The README's platform description, modular composition, and role as a gateway to subprojects are consistent, scoring 2. It points to external documentation and subprojects, but the supplied material contains no dependency inventory or availability guarantees, scoring 1. There is no evidence about error handling or failure messages, scoring 0.
The README identifies AI practitioners, platform administrators, and decision-makers, and mentions local, on-premises, and cloud settings, scoring 2. Capability boundaries are only described broadly through modularity and repository role, with no clear unsupported cases, scoring 1. No Agent trigger conditions or applicability rules are provided, scoring 0. Its Kubernetes, cross-environment, and AI/ML/HPC descriptions support some environment fit, scoring 2.
The README has sections for overview, mission, community, and repository role, scoring 2. The supplied material gives no installation instructions, scoring 0. The Kubeflow name is used consistently with the subproject description, scoring 2. There are no examples or FAQ, scoring 0. The README says this repository is primarily a gateway and that development happens in subproject repositories, partially stating scope limits, scoring 1. The LICENSE contains the Apache 2.0 terms, scoring 3. No version history or changelog is supplied, scoring 0. The README names working groups and governance committees as the maintenance structure, but gives no individual owners or update path, scoring 2.
The README helps readers understand the platform's purpose and this repository's role, but supplies no directly usable Agent output or operating procedure, scoring 1. It claims support for data, AI/ML, and HPC workloads across the AI lifecycle, but the sample does not demonstrate concrete reusable capabilities, scoring 1. The materials provide no deployment cost, operating burden, or benefit comparison, scoring 1.
The README explicitly states the platform description, governance, and repository role, but many capability claims have no implementation evidence in the sample, scoring 1. The sample contains only a README and license, with no code, tests, or other sources for corroboration, scoring 0. Mission and positioning are presented as project claims without measurement methods or validation results, so fact and inference are only partly distinguished, scoring 1.
- Not found in source: least-privilege scopingGrant only what the task needs: a dedicated account or read-only token, scoped to specific directories and repos.
- Not found in source: confirmation before actingTurn on (or add) a confirmation step before it acts, and try it in a sandbox or test environment before real data.
- Not found in source: data-flow disclosureWatch which external services it contacts (proxy or firewall logs) and keep sensitive data out until you know where it goes.
- Not found in source: sensitive-data handlingUse dedicated, low-privilege, revocable API keys — never production credentials — and keep secrets out of logs.
- Not found in source: dependency securityPin versions and run a dependency audit (npm audit, pip-audit) before installing; prefer running it in a container.
- Not found in source: disclosed external effectsEstablish which external systems it writes to, sends to or changes, and verify with test accounts or repos before production.
- Not found in source: rollback or recovery pathBack up first, or work on a git branch or snapshot, so its changes can be undone.
- This sample describes Kubeflow as a platform and subproject gateway, but provides no Agent implementation or execution-safety evidence; do not infer specific Agent capabilities or safeguards from it.
- Installation guidance, dependencies, tests, release history, and named maintenance owners are not included in the supplied material; verify them in the relevant subproject materials before use.
Recent releases
Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.
-
redirect Kubeflow Subprojects Releases
Release notes excerpt
Kubeflow brings together multiple open-source projects, each focused on a different stage of the [AI lifecycle](https://www.kubeflow.org/docs/started/architecture/#introducing-the-ai-lifecycle). If you want to install Kubeflow please refer to https://www.kubeflow.org/docs/started/installing-kubeflow/. Each project follows its own release cycle, please find links to the releases below: | Kubeflow Project | GitHub Releases | …View on GitHub ↗ -
Release notes excerpt
This release was cut for [Kubeflow Platform](https://www.kubeflow.org/docs/started/introduction/#what-is-kubeflow-platform) version [__1.10.0__](https://www.kubeflow.org/docs/releases/kubeflow-1.10/). __NOTE:__ the following release notes are __only__ for the components in the kubeflow/kubeflow repo. Significant Changes * feat: update notebook server images by @thesuperzapper in https://github.com/kubeflow/kubeflow/pull/7687 * chore: migrate docker images to ghcr by @ederign and @thesuperzapper in https://github.com/kubeflow/kubeflow/pull/7706 Features …View on GitHub ↗ -
Release notes excerpt
This release was cut for [Kubeflow Platform](https://www.kubeflow.org/docs/started/introduction/#what-is-kubeflow-platform) version [__1.9.1__](https://www.kubeflow.org/docs/releases/kubeflow-1.9/). __NOTE:__ the following release notes are __only__ for the components in the kubeflow/kubeflow repo. Fixes Central Dashboard * fix: logout button for oauth2-proxy @tarekabouzeid in https://github.com/kubeflow/kubeflow/pull/7654 Notebook Controller * fix: return notebook events error message @gigabyte132 in https://github.com/kubeflow/kubeflow/pull/7649 New Contributors …View on GitHub ↗
FAQ
Does Kubeflow require a particular cloud provider?
Can this repository directly install a complete Kubeflow environment?
Does Kubeflow require a paid license?
Does it natively integrate with ChatGPT, Claude, or their APIs?
Where should operators investigate failures?
Is Kubeflow safe to use?
Is Kubeflow still maintained?
Who makes Kubeflow?
What is Kubeflow's official website?
Thanks — we use this to improve the page.