Skip to main content

Google Vertex AI

This guide preserves the legacy integration pattern for Google Vertex AI. Review package versions and deployment commands before publishing.

Requirements​

  • Google Cloud project with billing enabled.
  • Vertex AI API enabled.
  • Python 3.10 or newer.
  • Mailtarget API access.
  • A verified company sending domain.

Do not send from public mailbox domains such as Gmail, Yahoo, or Outlook. Use a company domain or a dedicated sending subdomain that has passed Mailtarget domain verification.

Tool boundary​

Expose a server-side tool such as send_email(). The tool should validate input, read MAILTARGET_API_KEY from the server environment, and call the Transmission API.

import os
import requests

def send_email(to_email: str, subject: str, body_text: str) -> dict:
    response = requests.post(
        "https://transmission.mailtarget.co/v1/layang/transmissions",
        headers={
            "Authorization": f"Bearer {os.environ['MAILTARGET_API_KEY']}",
            "Content-Type": "application/json",
        },
        json={
            "subject": subject,
            "from": {
                "email": "agent@your-verified-domain.com",
                "name": "Your Company",
            },
            "to": [{"email": to_email}],
            "bodyText": body_text,
            "optionsAttributes": {
                "openTracking": True,
                "clickTracking": True,
                "transactional": True,
            },
        },
        timeout=10,
    )
    response.raise_for_status()
    return response.json()

Review checklist​

  • Confirm the current Google ADK or LangChain wiring for the target runtime.
  • Confirm whether gemini-2.0-flash is still the intended model.
  • Confirm deployment steps for Vertex AI Agent Engine.
  • Confirm whether domain creation and template creation tools should be exposed to agents or kept as admin-only functions.