Summary

The google-vertex hook alias exists but is a no-op — it routes through the same Gemini API OAuth flow with no Vertex AI-specific behavior. The Google provider needs proper Vertex AI support.

What's missing

• Location/region: Vertex AI requires a region to construct the endpoint URL ({region}-aiplatform.googleapis.com). No location field exists in the Google provider config or onboarding.
• Endpoint routing: The provider always hits generativelanguage.googleapis.com. Vertex AI uses {region}-aiplatform.googleapis.com/v1/projects/{project}/locations/{location}/publishers/google/models/{model}.
• Mode distinction: No way to distinguish between standard Gemini API mode and Vertex AI mode. The google-vertex alias should activate Vertex AI endpoint routing.

Requirements

• Add location config field to the Google provider (with GOOGLE_CLOUD_LOCATION / CLOUD_ML_REGION env var fallback, default us-central1)
• Construct Vertex AI endpoint URL when in Vertex mode
• Support Vertex AI OAuth (the existing OAuth flow should work, but token scopes may need https://www.googleapis.com/auth/cloud-platform)
• Onboarding flow should collect location when setting up via google-vertex
• Project ID (already collected) + location must be used for Vertex AI API calls

Reference

• Anthropic Vertex implementation for comparison: extensions/anthropic-vertex/region.ts
• Google OAuth files: extensions/google/oauth.*.ts
• Google provider entry: extensions/google/index.ts