Google AI Studio

Short verdict: Google AI Studio's Build mode turns a prompt into a full-stack web app or a native Android app, using the agent harness from Google Antigravity on Gemini models, and publishes web apps to Cloud Run in one click. It is free to start: Build mode costs nothing within Gemini's free-tier limits, and eligible accounts can publish 2 apps without a billing account. The catches are in the terms: outside the EEA, Switzerland and the UK, Google uses what you submit on free use to improve its products, and human reviewers may read it; the free Build-mode quota is not published; and Gemini 4 Argon needs a Google AI Ultra subscription in AI Studio. It suits prototypes, Gemini demos and Android experiments; confidential projects need a billing-enabled account or another builder.

How we reviewed it: We did not sign in to AI Studio or build an app. On 21 October 2026 we read the public AI Studio pages, the Gemini API documentation on Build mode, Android apps, deployment, billing and pricing, the Gemini API terms and Google's blog posts from October 2025 to October 2026. Google changes AI Studio often; check the Build mode docs before you rely on a limit.

What Google AI Studio costs and includes

What Build mode gives you

A full-stack app with a database and sign-in

Since March 2026 a web app gets a front end (React by default, or Angular or Next.js) and a Node.js server that installs npm packages, keeps API keys, including your Gemini key, in a server-side secrets manager and supports real-time multiplayer. When your app needs storage or accounts, the agent offers to provision Cloud Firestore and Firebase Authentication with Sign in with Google, and it can connect Gmail, Sheets, Docs, Drive, Calendar and Chat with OAuth handled for you.

Native Android apps without Android Studio

Since 19 May 2026, "Build an Android app" produces a Kotlin and Jetpack Compose project that runs in a cloud emulator in the browser, installs on a phone over ADB and publishes to the Google Play internal test track once you connect a Play developer account. Since 13 October 2026 Google Play has been rolling out installs of your own AI Studio apps with a standard Google account, no developer account needed, in the US and some other markets.

One-click publishing and a way out

Publish deploys a web app to Cloud Run with a custom address under ai.studio; beyond the 2 free Starter Tier apps, you need a billing-enabled Cloud project. GitHub sync works both ways, and one click exports the project, chat history and secrets to Google Antigravity for local work. Google is shutting down its other AI workspace, Firebase Studio, on 22 March 2027 and points its users to AI Studio or Antigravity.

Limits and terms to check before you build

• Free use feeds Google's products. The Gemini API terms say Google uses unpaid prompts and responses to improve its products, that human reviewers may read them, and that you should not submit sensitive, confidential or personal information. AI Studio's vibe coding FAQ says your data is not used to train models for others without permission; the terms are the binding text.

• Quotas you cannot see in advance. Google publishes no Build-mode quota, only per-model rate limits that you see after signing in. Apps you share or publish call Gemini with your key, so their use counts against your quota or bill.

• Starter Tier fine print. Google Workspace accounts and anyone with a current or past paid Google Cloud billing account are not eligible, support cannot override that, and the tier allows 2 services in one region. The $300 Google Cloud welcome credit cannot pay for Gemini API or AI Studio usage.

• Business use only. The terms say AI Studio is for developers building for professional or business purposes, not for consumer use, and apps offered to users in the EEA, Switzerland or the UK must run on paid services.

• Android limits. Android apps are client-side only (no secrets, Workspace APIs or multiplayer); Java, XML layouts, native code, Wear OS and Android TV are not supported, export is ZIP only, and Play publishing stops at internal testing.

Pros and cons

Pros

• Free to build within Gemini's free-tier limits, with 2 published apps free for eligible accounts.

• Full-stack output: Node.js server, secrets, npm packages and Firebase database and sign-in.

• Builds native Android apps with a browser emulator and Play test-track publishing.

• Code is portable through two-way GitHub sync, ZIP download and Antigravity export.

Cons

• Free use can train Google's models and be read by human reviewers outside the EEA, Switzerland and the UK.

• Free Build-mode quotas are not published, and Gemini 4 Argon needs Google AI Ultra in AI Studio.

• Google models only, and published apps bill Gemini calls to your key.

• No iOS apps, and Android output is client-side only.

How Google AI Studio compares

• Lovable builds and hosts full-stack web apps on a Supabase-based backend; Free gives 5 build credits a day and Pro costs $25 a month for 100 credits. Choose it for web apps with hosting included and no Google account.

• Bolt.new builds web apps and, through Expo, mobile apps with App Store and Google Play releases; Free includes 1 million tokens a month and Pro costs $25 a month. Choose it if you need an iPhone app.

• Google Antigravity is Google's desktop agent app and editor and the one-click export target for AI Studio projects; it is free with a weekly quota, and Google AI Pro and Ultra raise the limits. Choose it when a project outgrows the browser and you want to keep working locally.

Compare more vibe coding platforms in all AI app builders.

Who should use Google AI Studio

Build mode suits developers and makers who want to try a Gemini-powered idea, prototype a web app with a real database and sign-in, or put a personal Android app on their own phone without paying up front. It is a weaker fit for confidential code on a free account, public apps in Europe without a paid setup, teams that need real-time co-editing, which AI Studio lacks, and anyone who needs iOS apps or other vendors' models.

FAQ

Is Google AI Studio free?

Yes, Build mode is free within the Gemini API's free-tier rate limits, but Google does not publish how much building that allows. Heavier use needs a paid Gemini API key billed per token or a Google AI Pro ($19.99 a month) or Ultra (from $99.99 a month) subscription for higher limits, as of 21 October 2026.

Does Google use my AI Studio prompts and code for training?

On free use, yes, outside the EEA, Switzerland and the UK. The Gemini API terms say Google uses unpaid content to improve its products and machine learning and that human reviewers may read it. Accounts with access to a billing-enabled Cloud project, and Workspace enterprise accounts, get paid-service terms with no training.

Can Google AI Studio build mobile apps?

Yes, native Android apps in Kotlin and Jetpack Compose, which you can run in a browser emulator, install over ADB and publish to the Google Play internal test track. It does not build iOS apps, and Android apps are client-side only, without the server runtime web apps get.

Where are Google AI Studio apps hosted?

On Google Cloud Run, published in one click with an ai.studio address. Eligible accounts can publish 2 apps on the Google Cloud Starter Tier without a billing account; beyond that you need a billing-enabled Cloud project. You can also download the code as a ZIP or sync it to GitHub and host it elsewhere.