vertexai

What it is and what it does

The vertexai package is Google's official Python SDK for the Gemini Enterprise Agent Platform, a managed service for building and deploying AI agents. It wraps the underlying google-cloud-aiplatform library and provides high-level APIs for defining agents using the Agent Development Kit (ADK), running generative AI inference and evaluation, optimizing prompts, and deploying agents to a managed runtime. The package handles authentication, API communication, and lifecycle management for agents and generative AI workloads on Google Cloud.

Typical workflows include instantiating a client with your GCP project and location, defining agent functions or using ADK to create agents with tool bindings, running inference or evaluation on model outputs, and deploying agents to Agent Engine for production use. The SDK supports both local testing (async streaming) and remote deployment, making it suitable for developers building conversational agents, question-answering systems, and other AI-driven applications that need to integrate with Google Cloud infrastructure.

Use it for:

• Build and deploy multi-turn conversational agents with tool integration using the Agent Development Kit.
• Run batch inference and evaluation on generative AI model outputs using built-in metrics like exact_match and rouge_l_sum.
• Optimize prompts for better model performance using zero-shot or data-driven optimization methods.
• Manage agent sandboxes, sessions, and memory banks for stateful agent interactions.
• Evaluate and compare model responses against reference outputs in production workflows.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Python SDK for building, deploying, and managing AI agents and generative AI applications on Google's Gemini Enterprise Agent Platform (formerly Vertex AI).

Yes, if you are building AI agents or generative AI applications on Google Cloud. The package is actively maintained, has low install friction, carries a permissive license, and integrates deeply with Google's managed agent platform. Be aware that major API reorganization is planned (rebranding agent_engines to runtimes, relocating modules) before version 2.0.0; pin to < 2.0.0 if stability is critical. No known vulnerabilities.

Install

vertexai on PyPI

pip install vertexai

uv add vertexai

poetry add vertexai

Installing vertexai

Before you install

Low install friction with a single runtime dependency (google-cloud-aiplatform). Actively maintained with recent commits and steady releases; supports current Python versions (3.8+).

License in practice

Apache 2.0 permissive license allows commercial and private use with minimal restrictions; suitable for most production and proprietary projects.

Quickstart

pip install vertexai

import vertexai
client = vertexai.Client(project='my-project', location='us-central1')
inference_results = client.evals.run_inference(
    model="gemini-2.5-flash-preview-05-20",
    src=prompts_df
)

Requires Google Cloud credentials and a valid GCP project; API access to Gemini models must be enabled in your project.

Verify before relying

• Whether the upcoming rebranding (agent_engines → runtimes, module relocations) will occur before or after version 2.0.0 and what migration path is recommended.
• Performance characteristics and rate limits for concurrent agent deployments and inference calls.
• Compatibility guarantees with specific google-cloud-aiplatform versions beyond the single runtime dependency.

Package facts

Evidence: vertexai-1.71.1-py3-none-any.whl

Tags

More Artificial Intelligence packages