Applies to SUSE AI Factory

3 Quick start: copy an existing blueprint #

The fastest path to creating a blueprint is copying and adapting an existing one.

Download an example blueprint manifest YAML file.

wget https://github.com/SUSE/aif/blob/59be88cd2944b433720de3fe0c27042961f0f7e7/charts/aif-operator/files/blueprints/simple-chatbot-with-rag-1.0.0.yaml

Edit my-demo-ollama-1-0-0.yaml to configure your blueprint.

Parameter	Original value	New value
CR file name	`simple-chatbot-with-rag-1.0.0.yaml`	`my-demo-ollama-1-0-0.yaml`
`metadata.name`	`simple-chatbot-with-rag-1-0-0`	`my-demo-ollama-1-0-0` (use hyphens)
`labels[…/blueprint-name]`	`simple-chatbot-with-rag`	`my-demo-ollama`
`spec.displayName`	`Simple Chat Bot with RAG`	`My Demo (Ollama, CPU)`
`spec.description`	Original description text	Your customized description

# Run a server-side dry run first to validate schema
> kubectl apply --dry-run=server -f my-demo-ollama-1-0-0.yaml

# Apply the resource to register it
> kubectl apply -f my-demo-ollama-1-0-0.yaml
> kubectl get blueprints.ai-factory.suse.com

Deploy the workload from the SUSE AI Factory Web interface.
1. In the navigation menu, select Blueprints.
2. Select My Demo (Ollama, CPU) and click Create AIWorkload.
3. Choose the target namespace.
4. Monitor the pod deployment status.
```
> kubectl -n <NAMESPACE> get pods -w
```