CLI Commands Reference¶
The starter kit provides a unified CLI (./cli) for managing infrastructure, components, and examples. This page documents all available commands and their usage.
Core Commands¶
configure¶
Configure environment variables for the project.
Interactive prompts for:
REGION- AWS region (e.g.,us-west-2)EKS_CLUSTER_NAME- Name of the EKS clusterEKS_MODE- EKS mode:auto(default) orstandardDOMAIN- Domain name for ingress (optional)HF_TOKEN- Hugging Face user access token
Values are saved to .env.local.
Example:
./cli configure
# ✔ Enter value for REGION: us-west-2
# ✔ Enter value for EKS_CLUSTER_NAME: genai-on-eks
# ✔ Enter value for DOMAIN: example.com
demo-setup¶
Quick setup with infrastructure and curated demo components.
This command:
- Sets up required infrastructure using Terraform
- Deploys demo components specified in
config.json - Installs components in the correct dependency order
See Demo Walkthrough for details.
interactive-setup¶
Guided setup with component selection.
This command:
- Presents interactive menus for component selection
- Sets up required infrastructure using Terraform
- Installs selected components (order not guaranteed)
Example:
./cli interactive-setup
# ✔ Select AI Gateway components to install: litellm
# ✔ Select LLM Model components to install: vllm
# ✔ Select Embedding Model components to install: tei
Note
Some components may need to be re-installed after initial setup due to dependencies.
Component Management¶
Install a Component¶
Categories: - nvidia-platform - NVIDIA Platform components - ai-gateway - AI Gateway components - llm-model - LLM Model components - embedding-model - Embedding Model components - guardrail - Guardrail components - o11y - Observability components - gui-app - GUI App components - vector-database - Vector Database components - workflow-automation - Workflow Automation components - ai-agent - AI Agent components
Examples:
./cli ai-gateway litellm install
./cli llm-model vllm install
./cli embedding-model tei install
./cli o11y langfuse install
./cli gui-app openwebui install
./cli vector-database qdrant install
./cli ai-agent openclaw install
Uninstall a Component¶
Examples:
./cli ai-gateway litellm uninstall
./cli llm-model vllm uninstall
./cli embedding-model tei uninstall
Example Management¶
Install an Example¶
Categories: - mcp-server - MCP Server examples - strands-agents - Strands Agents examples - agno - Agno examples - openclaw - OpenClaw examples
Examples:
./cli mcp-server calculator install
./cli strands-agents calculator-agent install
./cli agno calculator-agent install
./cli openclaw doc-writer install
./cli openclaw devops-agent install
Uninstall an Example¶
Examples:
./cli mcp-server calculator uninstall
./cli strands-agents calculator-agent uninstall
./cli agno calculator-agent uninstall
./cli openclaw doc-writer uninstall
./cli openclaw devops-agent uninstall
Model Management¶
Configure Models¶
Configure which models should be deployed for a component.
Examples:
./cli llm-model vllm configure-models
./cli llm-model sglang configure-models
./cli embedding-model tei configure-models
This command presents an interactive menu to select models from the available list in config.json.
Update Models¶
Add and/or remove models for a component.
Examples:
Interactive prompts: 1. Select models to add (from available models) 2. Select models to remove (from currently deployed models)
Add Models¶
Add missing models without removing existing ones.
Examples:
This only adds new models - existing deployments are not touched.
Remove All Models¶
Remove all deployed models for a component.
Examples:
Warning
This will delete all model deployments for the specified component.
NVIDIA Platform Commands¶
Install NVIDIA Components¶
./cli nvidia-platform monitoring install # Prometheus + Grafana
./cli nvidia-platform gpu-operator install # NVIDIA GPU Operator
./cli nvidia-platform dynamo-platform install # Dynamo CRDs, Operator, etcd, NATS
./cli nvidia-platform dynamo-vllm install # Deploy model with vLLM
./cli nvidia-platform benchmark install # AIPerf concurrency sweep
./cli nvidia-platform aiconfigurator install # TP/PP recommendation + SLA deploy
See NVIDIA Platform Overview for detailed documentation.
Cleanup Commands¶
cleanup-infra¶
Destroy infrastructure using Terraform.
This command:
- Runs
terraform destroyto remove all AWS resources - Keeps component/example deployments (if any remain)
Warning
This will destroy the EKS cluster and all associated AWS resources.
cleanup-everything¶
Comprehensive cleanup of components, examples, and infrastructure.
This command:
- Attempts to uninstall all deployed examples
- Attempts to uninstall all deployed components
- Destroys infrastructure using Terraform
Warning
This is a destructive operation. Ensure you have backups of any important data.
Terraform Commands¶
terraform apply¶
Apply Terraform configuration.
This command:
- Initializes Terraform workspace based on
EKS_CLUSTER_NAME - Runs
terraform applywith auto-approve - Applies infrastructure changes
Use cases: - Update infrastructure after changing Terraform variables - Apply configuration changes from config.json - Re-apply infrastructure after manual modifications
Example:
# Update ECR Pull Through Cache settings
# Edit config.local.json to enable ECR Pull Through Cache
./cli terraform apply
Command Options¶
Common Patterns¶
All component/example commands follow the pattern:
Where: - <category> - Component or example category - <component|example> - Specific component or example name - <action> - install or uninstall
Finding Available Components¶
List all available components and examples in cli-menu.json:
cat cli-menu.json | jq '.componentCategories[].components[].name'
cat cli-menu.json | jq '.exampleCategories[].examples[].name'
Environment Variables¶
The CLI uses environment variables from:
.env(default values)config.json(default configuration).env.local(user overrides)config.local.json(user configuration)
Later sources override earlier ones.
Key Variables¶
| Variable | Description | Default |
|---|---|---|
REGION | AWS region | us-west-2 |
EKS_CLUSTER_NAME | EKS cluster name | genai-on-eks |
EKS_MODE | EKS mode | auto |
DOMAIN | Domain name for ingress | (empty) |
HF_TOKEN | Hugging Face token | (required) |
LITELLM_API_KEY | LiteLLM API key | (generated) |
OPENCLAW_GATEWAY_TOKEN | OpenClaw token | openclaw-gateway-token |
Troubleshooting¶
Command Not Found¶
Problem: ./cli: command not found
Solution: Ensure you're in the project root directory:
Permission Denied¶
Problem: ./cli: Permission denied
Solution: Make the CLI executable:
Component Installation Fails¶
Problem: Component fails to install
Solution: 1. Check prerequisites are installed (kubectl, helm, docker) 2. Verify AWS credentials are configured 3. Check cluster is accessible: kubectl get nodes 4. Review component logs: kubectl logs -n <namespace> -l app=<component>
Terraform Errors¶
Problem: Terraform commands fail
Solution: 1. Ensure Terraform is installed: terraform version 2. Check AWS credentials: aws sts get-caller-identity 3. Review Terraform state: cd terraform && terraform show 4. Clean and re-initialize: cd terraform && rm -rf .terraform && terraform init
Examples¶
Complete Setup Flow¶
# 1. Configure environment
./cli configure
# 2. Deploy infrastructure and demo components
./cli demo-setup
# 3. Install additional components
./cli o11y langfuse install
./cli vector-database qdrant install
# 4. Install examples
./cli mcp-server calculator install
./cli agno calculator-agent install
# 5. Configure models
./cli llm-model vllm add-models
# 6. Access services
kubectl get ingress --all-namespaces
Selective Installation¶
# Infrastructure only
./cli terraform apply
# Core components
./cli ai-gateway litellm install
./cli llm-model vllm install
./cli gui-app openwebui install
# Configure models
./cli llm-model vllm configure-models
# Install specific examples
./cli openclaw doc-writer install
Cleanup¶
# Remove specific components
./cli llm-model vllm uninstall
./cli ai-gateway litellm uninstall
# Complete cleanup
./cli cleanup-everything