Cloud Run
Cloud Run is the compute layer. Every Google AI Studio deployment creates a Cloud Run service that handles HTTP traffic. Under the Starter Tier, you can deploy up to two active web applications at a time per Google Account. Cloud Run services scale automatically based on incoming traffic and scale down to zero when idle, meaning your prototypes don’t consume resources when not in use. They run in a single region that is locked in when you first provision your Starter Tier environment.
Firebase Authentication
If your app needs user login, the Starter Tier includes Firebase Authentication with Google Sign-In preconfigured. The AI agent in Google AI Studio can detect when your prompt implies user identity (for example, “build a shared to-do list”) and will offer to enable auth automatically.
If your application builds on Google Workspace integrations, this sign-in flow simplifies credentials. Once a user logs in, your application can request OAuth access scopes to securely interact with their Gmail, Docs, Calendar, or Sheets data, making it straightforward to prototype internal tools like summarizers or inbox sorters.
Cloud Firestore
Cloud Firestore is a database service that handles NoSQL data storage. The Google AI Studio agent can provision it automatically when your prompt implies the need for structured data storage. The AI agent generates the client-side sync code (typically a /src/lib/firebase.ts file), and drafts application-appropriate Firebase Security Rules (for example, utilizing request.auth.uid to restrict document access to the authenticated creator).
If you hit a “Missing or insufficient permissions” error, you can click “Fix error” in Google AI Studio, and the agent will rewrite the security rules to match your updated app logic. It’s worth reviewing these security rules manually before sharing your app broadly, though. AI-generated security rules are a starting point, not a guarantee.
All Firestore databases created by the Google AI Studio agent share a usage quota (more on that in the limits section below).
Cloud SQL for PostgreSQL Developer edition
When you need relational data with proper schemas, joins, and ACID compliance, the Starter Tier provisions Cloud SQL for PostgreSQL Developer edition, designed to work seamlessly with AI Studio agent. The developer edition enables instant provisioning and scale to 0, which enables fast and low cost developer experience. You also get the full power of open source PostgreSQL with capabilities like pgvector, so you can build semantic search or RAG applications without bolting on a separate vector database.
As you iterate on your application using prompts, Google AI Studio agent will automatically generate the required schema and migrate the schema, as you move through building and publishing your application.
