adk-cli 🚀

A modern CLI tool to scaffold, develop, and deploy Google ADK (Agent Development Kit) projects in Python.

✨ Features

• ⚡ Quickly scaffold a new ADK agent project with best practices
• 🛠️ Generates agent, runner, .env, Streamlit chat, tests, and deployment files
• 🚄 Supports uv for fast dependency management
• ☁️ Built-in deploy command for Google Cloud Run
• 🧹 Pre-configured with ruff for linting
• 🖥️ Installable as a CLI tool: adk-cli

📦 Installation

From PyPI (recommended)

Once published, you can install globally:

pip install adk-cli

Local development

Clone this repo and install locally:

pip install .

Or, for development mode:

pip install -e .

🚀 Usage

Scaffold a new project

adk-cli new

You will be prompted for project name, GCP project, and API key. A new folder will be created with all necessary files.

Deploy to Google Cloud Run

From within your generated project:

adk-cli deploy

Run the agent locally

python runner.py

Run the Streamlit chat UI

streamlit run streamlit_app/app.py

🗂️ Project Structure

myproject/
├── agent.py
├── runner.py
├── __init__.py
├── .env.example
├── pyproject.toml
├── README.md
├── streamlit_app/
│   └── app.py
├── tests/
│   └── test_agent.py
├── deploy/
│   ├── Dockerfile
│   └── cloudrun.yaml
└── ...

🛠️ Development

• The CLI is implemented in the adk_cli Python package.
• Entry point is defined in pyproject.toml as adk-cli.
• Uses Typer for CLI logic.
• To test locally, use pip install -e . and run adk-cli from anywhere.

🤝 Contributing

PRs and issues welcome! Please lint with ruff and follow best practices for Python packaging and CLI design.

�� License

Apache 2.0