Skip to main content
Google Vertex AI gives you the same Gemini models as Google AI Studio but with enterprise-grade features: VPC-SC, CMEK, data residency controls, and IAM-based access.

​
Setup

Vertex AI uses Application Default Credentials (ADC) instead of an API key. 
export GOOGLE_CLOUD_PROJECT=my-project-id
gcloud auth application-default login

​
Config

{
  "providers": {
    "google-vertex": {
      "baseUrl": "https://us-central1-aiplatform.googleapis.com"
    }
  }
}

​
Use it

{
  "agents": [
    { "name": "coder", "model": "google-vertex:gemini-2.5-pro" }
  ]
}

​
Models

Same models as Google Gemini — gemini-2.5-pro, gemini-2.5-flash, gemini-2.0-flash.

​
Features

FeatureSupported
StreamingYes
Tool useYes
Vision (images)Yes
ReasoningYes (2.5 Pro, 2.5 Flash)
VPC-SCYes
CMEKYes
Data residencyYes

​
Provider Details

Provider IDgoogle-vertex
Env variableGOOGLE_CLOUD_PROJECT + Application Default Credentials
API typeVertex AI API

​
When to Use Vertex vs. Google AI Studio

Google AI StudioVertex AI
AuthAPI keyIAM / ADC
ComplianceNoneVPC-SC, CMEK, HIPAA
Data residencyNo controlRegion-specific
BillingPay-as-you-goGCP billing
Best forPersonal/startupEnterprise

​
Notes

  • If you’re already running on GCP, Vertex AI avoids the need for a separate API key — your service account handles authentication.
  • Pricing is the same as Google AI Studio for most models.
  • Make sure GOOGLE_CLOUD_PROJECT is set and your ADC credentials have the aiplatform.endpoints.predict permission.