Jozu Hub Overview

What is Jozu Hub?

Jozu Hub is a self-hosted or SaaS platform for packaging, securing, and deploying AI/ML models using the tools you already use - like Jupyter, MLflow, or Weights & Biases. It helps data scientists and ML engineers create open source KitOps ModelKits, which are secure, portable, and traceable model packages, built directly from common tools like Jupyter, MLflow, or Weights & Biases.

With Jozu, models become first-class assets - versioned, signed, auditable, and ready for production deployment with confidence.

Jozu Hub can be used as SaaS, installed on-premises, or deployed to a private environment so it is 100% under organizational control.

It provides greater security, privacy, and control than public registries like Hugging Face, making it ideal for organizations in any regulated industry or those that value data privacy and cleanliness in their AI/ML projects.

How Jozu Hub relates to KitOps

KitOps is the open source standard that defines how Jozu packages AI/ML models. Jozu Hub builds on KitOps by providing model security, governance, and deployment automation for real-world teams.

Try It Out

After you've signed up for your free Jozu Hub account, you can pack and push your AI project with the KitOps CLI.

brew install kitops
kit init .
kit pack . -t myaiproj:v1.0
kit push myaiproj:v1.0 jozu.ml/[your-username]/myaiproj:v1.0

There's more information in our the step-by-step guide.

Why It’s Needed

Jozu addresses critical gaps in:

• Security - automatically scans for vulnerabilities and enforces signed attestations before deployment.
• Governance - every model version is tracked with lineage and metadata, enabling reproducibility and audit readiness.
• Deployment integrity - containers are auto-generated for each model, and deployments can be blocked if they lack required controls.

As regulatory and operational risks increase, Jozu enables safe, controlled AI deployment without slowing teams down.

In contrast, most AI/ML teams work in fragmented environments: Models are often scattered across systems like Git, S3, and Hugging Face, lacking consistent versioning, traceability, or auditability.

KitOps ModelKits plus the OCI registry gives organizations the same controls, security, and usability for their AI projects that git repositories do for their code.

Common Use Cases

Jozu Hub is designed to be flexible and self-hosted, making it suitable for various enterprise and regulated environments. It’s especially valuable for:

• 🧪 ML teams – Needing security and compliance without changing tools or workflows
• ⚙️ MLOps / DevOps – Needing versioned, policy-driven model deployments (like GitOps)

Jozu is usually overkill for hobbyists or anyone prototyping with public AI services.

Key Capabilities

With Jozu Hub, teams can:

1. Create a private, internal catalog of AI/ML models, stored in your own container registry and governed with your RBAC, auth, and policy controls.
2. Import models from sources like Hugging Face, apply security scanning, and wrap them with reproducible metadata and signatures.
3. Auto-generate containers from ModelKits, with deployment artifacts for Kubernetes and other orchestration platforms.
4. Block unsafe deployments by enforcing signed policies at runtime.

All while using the tools you already know - no data leaves your environment, and there’s no vendor lock-in.

Get started with the Jozu Hub

What are KitOps ModelKits?

KitOps is an open source project governed by the Cloud Native Computing Foundation (CNCF) - the same foundation that governs Kubernetes, OpenTelemetry, and Prometheus.

KitOps ModelKits revolutionize the way AI/ML artifacts are shared and managed throughout the lifecycle of AI/ML projects. ModelKits:

• Encapsulate datasets, code, configurations, and models into a single, standardized unit.
• Can be stored in your existing container registry - no new storage systems, no additional security approvals.
• Are based on the existing OCI standard which is the same standard containers use, so they're secure and tool-compliant.

Jozu Hub’s ModelKits are a great starting point for enterprise AI/ML projects. The Discover page lists the most popular, trending, and newest models, or you can use the search bar to find what you need. If you don't see a model you'd like you can import it from Hugging Face.

You can learn more about ModelKits in the documentation for the open source KitOps project.

Get started with the Jozu Hub