Are you an LLM? You can read better optimized documentation at /docs/understanding-jozu-hub/modelkit-containers.md for this page in Markdown format
Jozu Hub Runtime Inference Containers (RIC)
Jozu Rapid Inference Containers (RICs) are optimized, ready-to-run OCI-compatible containers that streamline AI model deployment. Each RIC is automatically built from a ModelKit, combining a runtime environment with the model, code, and data — no manual containerization required.
🔧 How It Works
Every ModelKit includes a manifest with one or more layers (e.g., model, datasets, code). These layers follow the Open Container Initiative (OCI) specification, just like Docker images.
Jozu combines these two layer types:
- Base container layers — OS, runtime, libraries
- ModelKit layers — model weights, metadata, scripts
At deploy time (e.g., via docker run or in Kubernetes), the container engine assembles both. Because both are OCI-compatible, the final image preserves immutability, reproducibility, and layer reuse — just like traditional containers.
🔒 Immutable by Design
ModelKits are content-addressable so each has a unique digest that guarantees integrity. When you build or pull a RIC:
- The model and data inside are byte-for-byte identical to the ModelKit
- You can verify that the deployed container uses the exact ModelKit you expect
This ensures safe, auditable, and tamper-proof deployment.
⚠️ Limitations
RICs use ModelKit layers as-is. This introduces some constraints:
- ModelKit contents are mounted into the container root directory
- Files are owned by root, with original permissions
- No file transformations or relocations are performed
While we can work around these limitations by designing containers to work with ModelKits directly, you may run into compatibility issues on certain platforms. If you encounter issues or have any feedback, please email us at [email protected].