What would you build in Google AI Studio?

Use Build to describe a web app or a native Android app. For a web app that helps a library draft event blurbs, say what the visitor enters, what Gemini should suggest and how the visitor can edit the result. Native Android uses a separate build and test path.

Who pays when someone uses the AI feature?

When you share a Gemini-powered web app, Google says other people’s calls count toward your usage limits and may incur costs on paid models. The app can keep its Gemini key on the server. If you also need sign-in or saved records, decide whether to connect Firebase Authentication and Firestore and what each user should be able to see.

When does a web preview become a public app?

Publishing from Build hosts the web app on Cloud Run and gives it a public address. Eligible accounts can publish up to two apps on the Google Cloud Starter Tier without setting up a billing account. Check eligibility and the hosting and model limits before sharing widely.

What needs to work after you move the code?

Download the web project files or sync the code with GitHub. Your new host still needs to run the app’s server code and supply its private settings. List any data, sign-in and model services separately before deciding what you can keep or move.

What happens if the event blurb has a wrong date?

Sketch the result for a fictional event on June 12 and decide how someone spots and corrects a draft that says June 21.

Write the facts you know
List the made-up event title and date separately from the text Gemini would draft.
Show the correction
Draw an editable draft and leave sending or publishing as a distinct action.
Count a retry
Mark the initial generation and a second generation as two model requests. Show a useful message if the second request is unavailable.