create

The Agent Starter Pack provides CLI commands to help you create and manage AI agent projects.

Create Command

The create command helps you create new GCP-based AI agent projects from templates.

agent-starter-pack create PROJECT_NAME [OPTIONS]

Arguments

• PROJECT_NAME: Name for your new agent project directory and base for resource naming. Note: This name will be converted to lowercase and must be 26 characters or less.

Options

The following options will be prompted interactively if not provided via the command line:

• --agent, -a: Agent name or number to use. Lists available agents if omitted.
• --deployment-target, -d: Deployment target (agent_engine or cloud_run). Prompts if omitted.
• --datastore, -ds: Datastore for RAG agents (vertex_ai_search or vertex_ai_vector_search). Prompted if --include-data-ingestion is specified, or if the selected agent (e.g., agentic_rag) requires data ingestion, and this option is omitted.
• --region: GCP region for deployment (default: us-central1). Prompts for confirmation if not specified and --auto-approve is not used.

GCP account and project ID are detected automatically (using your active gcloud config settings). You will be prompted to confirm or change them unless --auto-approve is used.

Additional options:

• --include-data-ingestion, -i: Include data ingestion pipeline components (required by some agents like agentic_rag, which enable this automatically). If specified manually without --datastore, you will be prompted to select one.
• --debug: Enable debug logging.
• --output-dir, -o: Output directory for the project (default: current directory).
• --auto-approve: Skip interactive confirmation prompts for GCP credentials and region.
• --skip-checks: Skip verification checks for uv installation, GCP authentication, and Vertex AI connection.

Example Usage

# Create a new project interactively
agent-starter-pack create my-agent-project

# Create with specific agent, deployment target, region, and include data ingestion with Vertex AI Search
agent-starter-pack create my-agent-project -a agentic_rag -d cloud_run --region europe-west1 -i -ds vertex_ai_search

# Create without interactive prompts (uses detected GCP credentials)
agent-starter-pack create my-other-agent -a chat_agent -d agent_engine --auto-approve

# Create in a specific output directory
agent-starter-pack create my-specific-loc-agent -o ./my-agents/