gemini-api

Use when using the Gemini API on Agent Platform with Google Gen AI SDK.

SKILL.md

GitHub

Gemini API in Agent Platform

Access Google's most advanced AI models for enterprise use cases using the Gemini API in Agent Platform (formerly Vertex AI).

SDK Installation

Language Package Install Command
Python google-genai pip install google-genai
JS/TS @google/genai npm install @google/genai
Go google.golang.org/genai go get google.golang.org/genai
Java com.google.genai:google-genai Add to build.gradle or pom.xml
C# Google.GenAI dotnet add package Google.GenAI

Authentication

export GOOGLE_CLOUD_PROJECT='your-project-id'
export GOOGLE_CLOUD_LOCATION='global'
export GOOGLE_GENAI_USE_ENTERPRISE=true

Code Example: Python

from google import genai

client = genai.Client()
response = client.models.generate_content(
    model="gemini-3.6-flash",
    contents="Explain quantum computing",
)
print(response.text)

Code Example: JavaScript/TypeScript

import { GoogleGenAI } from "@google/genai";
const ai = new GoogleGenAI({ enterprise: { project: "your-project-id", location: "global" } });
const response = await ai.models.generateContent({
    model: "gemini-3.6-flash",
    contents: "Explain quantum computing"
});
console.log(response.text);

Recommended Models

  • gemini-3.1-pro-preview — Complex reasoning, coding, research (1M tokens)
  • gemini-3.6-flash — Fast, balanced performance, multimodal (1M tokens)
  • gemini-3.5-flash-lite — High-frequency, lightweight tasks (1M tokens)

Common Pitfalls

  • Legacy SDKs: Do NOT use google-cloud-aiplatform, @google-cloud/vertexai, or google-generativeai — they are deprecated
  • Model naming: Use correct model names (e.g., gemini-3.6-flash, not legacy gemini-pro)
  • Enterprise flag: Set GOOGLE_GENAI_USE_ENTERPRISE=true for Agent Platform access

Verification Checklist

  • Gen AI SDK installed for the target language
  • Environment variables configured (GOOGLE_CLOUD_PROJECT, GOOGLE_CLOUD_LOCATION)
  • API enabled: gcloud services enable aiplatform.googleapis.com
  • Basic generate_content call succeeds