Truly One-Click: Full-Stack Deployment from Google AI Studio to Cloud Run [Next '26]

Introduction

Hello, this is Toriumi.
At Google Cloud Next '26 (April 22–24, 2026), the deployment from Google AI Studio to Cloud Run received a major upgrade.
https://cloud.google.com/blog/products/serverless/whats-new-for-cloud-run-at-next26?hl=en

While "one-click deployment" existed previously, the essential change in this update is that you can now deploy a full-stack app (Firestore + Authentication + server-side code) in a single click.

I have tested it out, and I will write honestly about the process, including the parts where I got stuck.

What is "New"?

Here is a summary comparing Before and After.

	Before	Next '26 onwards (Now)
Deployable items	Simple frontend only	Full-stack (Server-side + Firestore + Auth)
Firestore integration	Manual configuration required	Automatic configuration with one click
User authentication	Manual configuration required	Automatic configuration with one click
Streaming support	Not supported in one click	Supported
Status	Preview	GA (General Availability)

The "deploy button" itself is not new. The core essence of this update is that a full-stack including Firestore and Auth has become one-click.

What I built

To best demonstrate the benefits of the full-stack configuration, I built an AI brainstorming board.

• Enter a theme → Gemini generates 6 idea cards
• Session (theme + cards) is automatically saved to Firestore
• Manage history per user with Google Sign-In

The point is that the "necessity" of using Firestore and Auth is clear.

Steps

Prerequisites

• Google Cloud account (with billing enabled)
• Browser only. No code required.

Step 1: Generate the app in AI Studio

Open Google AI Studio and click "Build" in the left menu.

Clarifying Names

There are several similar services, which can be confusing. What we are using this time is Google AI Studio at aistudio.google.com. It is not Firebase Studio or Agent Studio.

I entered the following prompt:

Create a full-stack AI brainstorming board app with the following features:

## Core Features
- User enters a theme or topic (e.g. "new app ideas", "marketing strategies")
- Gemini generates 6 creative idea cards for that theme
- Each card shows: a title, a 2-sentence description, and an emoji icon
- Cards are displayed in a responsive grid layout (2-3 columns)

## Data Persistence (use Firestore)
- Each brainstorming session (theme + generated cards) is saved automatically
- Users can view their past sessions in a sidebar or history list
- Users can delete sessions they no longer need

## Authentication (use Firebase Auth)
- Google Sign-In only
- Each user sees only their own sessions

## UI Design
- Clean, modern design with a white card on a soft gradient background
- Each idea card has a distinct pastel color accent
- Smooth fade-in animation when cards appear
- Mobile-friendly layout

## Tech
- Use Gemini API (gemini-2.0-flash) server-side for idea generation
- Return ideas as structured JSON from the server

By specifying server-side for idea generation, the Gemini API key is managed on the server side (it is not exposed to the client).

It generates the code in 2–3 minutes.

Step 2: Deploy to Cloud Run

Click the 🚀(Publish) button in the top right → Select project → "Deploy App"

...or so I would like to say, but I got stuck here.

Stuck Point ①: API key is not linked to the project

I got this error when trying to deploy:

This app uses Gemini API but no project is associated with your Gemini API key.
In order to publish this app, please go to the Secrets tab and associate
your Gemini API key with a project.

Cause: There are two types of Gemini API keys.

Type	Usage
Personal key (not project-linked)	Local development/prototyping only
Project-linked key	Required for Cloud Run deployment

Steps to resolve:

1. Open aistudio.google.com/apikey
   
2. Select "Create API key in existing project"
3. Choose the Google Cloud project to use
4. Copy the created key
5. Open the "Secrets" tab in the AI Studio Build screen
   
6. Update the GEMINI_API_KEY value with the new key

After resolving the API key issue, click 🚀(Publish) → Deploy App → Wait a few minutes, and a URL will be issued.

A service was also created in Cloud Run. Incidentally, although I did nothing regarding Firestore, it was created as well.

About Costs (I will be honest)

I thought it was "almost free," but there was a point to note.

Item	Reality
AI Studio (Browser)	Free
Gemini API	Free up to 1,500 requests per day for Flash models. Charged if exceeded
Cloud Run	⚠️ Depends on the region
Firestore	Within the free tier for validation levels

The free tier for Cloud Run is limited to US regions (us-central1, us-east1, us-west1).

In this deployment, it was automatically selected to asia-south1 (India). Since the same region was selected the second time, I assume it chose a nearby region. (I don't think there was a button to choose one.)

Countermeasure: It is safe to delete the service from the Cloud Run console once the article writing and validation are finished. If you delete it today, the cost will be negligible (a few cents or less).

Conclusion

Perspective	Evaluation
Full-stack one-click deployment	✅ Really worked
Stuck point	Project linking of API key (only a one-time effort)
Cost	⚠️ Stays within the free tier if US region is selected
For articles or verification use	◎ Highly usable

With the Next '26 update, the flow of "Building an app with AI → Publishing it including Auth & DB" has effectively become a no-code process. The hurdle from Vibe Coding to production release has definitely been lowered.

I think I will be using this in situations where I "want to show a prototype to someone immediately."

Reference Links

• Google AI Studio
• New features for Cloud Run announced at Next '26 | Google Cloud Blog
• Deploying from AI Studio to Cloud Run | Google Cloud Blog
• Cloud Run Pricing | Google Cloud
• Deploy from AI Studio to Cloud Run | Google Codelabs