🚀 Getting Started

This guide quickly walks you through setting up your first agent project.

Want zero setup? 👉 Try in Firebase Studio or in Cloud Shell

Prerequisites

Ensure you have:

1. Python 3.10+
2. Google Cloud SDK: Install Guide.
3. Terraform: Install Guide.

1. Install the Starter Pack

# Create and activate a Python virtual environment (Recommended)
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install the package
pip install agent-starter-pack

2. Create Your Agent Project

Run the create command and follow the prompts:

agent-starter-pack create my-awesome-agent

This command:

• Lets you choose an agent template (e.g., adk_base, agentic_rag).
• Lets you select a deployment target (e.g., cloud_run, agent_engine).
• Generates a complete project structure (backend, optional frontend, deployment infra).

Examples:

# Create a RAG agent for Cloud Run (select options when prompted)
agent-starter-pack create my-rag-agent

# Create a base ADK agent for Agent Engine directly
agent-starter-pack create my-adk-agent -a adk_base -d agent_engine

3. Explore and Run Locally

cd my-awesome-agent && make install && make playground

Inside your new project directory (my-awesome-agent), you'll find:

• app/: Backend agent code.
• deployment/: Terraform infrastructure code.
• tests/: Unit and integration tests for your agent.
• notebooks/: Jupyter notebooks for getting started with evaluation.
• frontend/: (If applicable) Web UI for interacting with your agent.
• README.md: Project-specific instructions for running locally and deploying.

➡️ Follow the instructions in your new project's README.md to run it locally.

Next Steps

You're ready to go!

• Customize: Modify the code in app/ to edit prompts, configure tools, adjust API endpoints, customize business logic, extend functionality!
• Deploy in minutes: Use the setup-cicd command.
• Customized the deployment: Use Terraform in deployment/ to deploy to Google Cloud or follow the deployment guide (Deployment Guide).
• Add Data (RAG): Configure Data Ingestion.
• Monitor: Explore Observability features.