About the Google Vertex connector

Integration Service connector for Google Vertex in Automation Cloud, enabling API access to Google's unified AI platform for building, deploying, and scaling ML models.

Google Vertex AI is a unified artificial intelligence platform that offers all of Google's cloud services under one roof. With Vertex AI, you can build ML models or deploy and scale them easily using pre-trained and custom tooling. The UiPath Connector for Google Vertex offers API access to the PaLM2 model.

Authentication

Before automating processes, you need to establish a connection with your Google Vertex instance. Step-by-step instructions are available in the Google Vertex authentication page.

Events

The Google Vertex connector does not currently support events.

Activity package

You can use the connection to enable the Google Vertex activities in UiPath® Studio to build automations that interact with your Google Vertex data or execute Vertex AI agents in agentic processes.

Vertex AI agent supported features

This connector does not support Vertex Conversational Agents. (e.g. Dialogflow CX). See Google Vertex AI agent connector features table for a list of features supported with Vertex AI agents in Maestro.

Agent deployment requirement

To run a Vertex AI agent as a service task in Maestro, the agent must be deployed with AdkApp configured. The connector calls create_session, which Vertex AI exposes only when the agent is wrapped in AdkApp. AdkApp is also what makes user_id and session state accepted inputs.

If the agent is not configured with AdkApp, its input parameter schema does not appear in the service task - selecting Refresh schema on the connection does not populate it - and running the agent fails with a 400 error reporting that session.user_id is required for a session.

Agents created through the Agent Studio UI are not wrapped in AdkApp by default. For details, refer to Create an ADK agent in the Google Cloud documentation.