Documentation for engineers integrating AI Cluster Runtime (AICR) into CI/CD pipelines, GitOps workflows, or larger platforms.
This section is for integrators who:
- Build automation pipelines using the AICR API
- Deploy and operate the AICR API server in Kubernetes
- Create custom recipes for their environments
- Integrate AICR into GitOps workflows (ArgoCD, Flux)
| Document | Description |
|---|---|
| Automation | CI/CD integration patterns for GitHub Actions, GitLab CI, Jenkins, and Terraform |
| Data Flow | Understanding snapshots, recipes, validation, and bundles data transformations |
| Kubernetes Deployment | Self-hosted API server deployment with Kubernetes manifests |
| EKS Dynamo Networking | Security group prerequisites for Dynamo overlays on EKS |
| Recipe Development | Creating and modifying recipe metadata for custom environments |
| Validator Extension | Adding custom validators and overriding embedded ones via --data |
# Deploy API server to Kubernetes
kubectl apply -k https://github.com/NVIDIA/aicr/deploy/aicrd
# Generate recipe via API
curl "http://aicrd.aicr.svc/v1/recipe?service=eks&accelerator=h100"# GitHub Actions example
- name: Generate recipe
run: |
curl -s "http://aicrd.aicr.svc/v1/recipe?service=eks&accelerator=h100" \
-o recipe.json
- name: Generate bundles
run: |
curl -X POST "http://aicrd.aicr.svc/v1/bundle?bundlers=gpu-operator" \
-H "Content-Type: application/json" \
-d @recipe.json \
-o bundles.zip- Users: See User Documentation for CLI usage and installation
- Contributors: See Contributor Documentation for architecture and development guides