This is a simple agent built using Google Agent Development Kit (ADK). It is designed to be as simple as possible to help you get familiar with ADK tools, deployment to Vertex AI Agent Engine, and accessing agents through Gemini Enterprise.

The agent uses the Gemini model and Google Search grounding to provide wacky and interesting fun facts about any topic you provide.

• fun_facts/agent.py: Contains the root_agent definition and the App configuration.
• pyproject.toml: Lists the Python dependencies and project metadata.
• .env.example: Template for required environment variables.

• Python 3.10+
• Google Cloud Project with the Vertex AI API enabled.
• gcloud CLI installed and authenticated (gcloud auth application-default login).
• ADK CLI installed (pip install google-adk).
• uv installed (Install uv).

1. Install Dependencies: From this directory, run:

   uv sync

2. Configure Environment: Copy the example environment file and fill in your Google Cloud Project ID:

   cp .env.example fun_facts/.env

   Edit .env to set your GOOGLE_CLOUD_PROJECT.

You can interact with your agent locally using the ADK CLI tools:

• Terminal: Run the agent directly in your terminal.

  adk run fun_facts

• Web UI: Launch a local web interface to chat with the agent and inspect its execution trace.

  adk web

Deploying your agent to Google Cloud allows it to be used as a managed service.

NOTE: For this agent, --region must be passed explicitly because the Gemini model requires the global endpoint, but Agent Engine must be deployed to a regional endpoint (e.g. us-central1).

adk deploy agent_engine fun_facts --region="us-central1"

This command will package your agent, upload it to Vertex AI, and create an Agent Engine resource.

The resource name will look something like this:

projects/PROJECT_NUMBER/locations/us-central1/reasoningEngines/AGENT_ENGINE_ID

You can view Agent Engine resources in the Cloud Console here: https://console.cloud.google.com/vertex-ai/agents/agent-engines.

Once deployed to Agent Engine, your agent can be made available to users in your organization through Gemini Enterprise:

• Follow the steps in the documentation to Register and manage ADK agents hosted on Vertex AI Agent Engine