Google Vertex AI
Leverage your existing Google Cloud Platform infrastructure and compliance certifications to access Claude.
What You Get
- Use existing GCP contracts and committed spend
- Compliance certifications already in place
- VPC Service Controls for network security
- Integrated billing with other GCP services
What You Can Do
- Access Claude 3.5 Sonnet and other supported models
- Route requests through your GCP project
- Apply IAM policies for access control
- Monitor usage in GCP billing
Prerequisites
- Google Cloud project with Vertex AI enabled
- Service account with Vertex AI permissions
- Claude models enabled in your region
Required Permissions
Your service account needs these roles:
roles/aiplatform.user- To invoke modelsroles/serviceusage.serviceUsageConsumer- To use the API
Setup
- Navigate to Settings → General → API Configuration
- Select Google Vertex AI as your provider
- Enter your configuration:
- Project ID
- Region (e.g.,
us-central1) - Service account credentials (JSON)
- Click Test Connection
- Save your configuration
Example Uses
Once configured, Sasha uses Vertex AI automatically:
"Summarize these meeting notes"
"Help me understand this error log"
"Draft an email to the team about the project update"