At Google I/O 2026, Google announced that its web-based Google AI Studio can now create native Android apps — a process the company says shrinks weeks of setup and coding down to minutes. The new flow generates Kotlin apps using Google's Jetpack Compose toolkit and can tie those apps directly to phone hardware such as GPS, Bluetooth and NFC.
Google said the capabilities are aimed at both seasoned developers and first-time creators, and it showed a preview of the experience: users build in a browser, interact with their app inside an embedded Android Emulator, then install a working copy on a phone over a USB cable using the Android Debug Bridge, or adb. For developers, Google said AI Studio can automatically create the app record, package the bundle and upload it to an internal testing track in Google Play Console.
The tooling also supports handoff. Projects can be downloaded as a zip file for local work in Android Studio or exported directly to GitHub, Google said. Behind the scenes the apps are standard Kotlin projects built on Jetpack Compose, which means the code should be portable for teams that want to take a generated prototype and mature it into a production release.
Google presented the new web builder as an extension of past moves to put generative models at the center of coding workflows. The company previously added Gemini-powered code assistance to the desktop version of Android Studio; bringing native app creation into the browser is the next step, Google said. Google is also planning mobile access to AI Studio and the ability to publish apps for family and friends in time, the company said.
There are limits, and those limits supply the story's friction. For now the resulting creations are explicitly for personal use only; publishing beyond family and friends remains on the roadmap. Google also said it plans to add Firebase integrations — including Firestore, Firebase Auth and Firebase App Check — but those server and authentication features are not yet live. That gap matters because many real-world apps rely on backend services and user management before they can be shared broadly.
The company is rolling out companion features to help users discover and deploy what they make. A new Ask Play AI overlay will let people find apps by having natural conversations with AI inside the Play Store, Google said, and Gemini will begin surfacing apps in users' conversations with the assistant. That app-surfacing rollout will begin in the weeks ahead across Gemini on the web and on Android, Google said.
Practical details underscore how fast the prototype loop can be. Google said creators can preview and interact with an app as it is being built in the browser emulator, then install it over USB with adb. For testers and small teams, AI Studio can push the packaged app to an internal testing track in Google Play Console, skipping much of the manual setup that can turn a concept into a working test build.
Google framed the feature set as lowering the barrier to building phone-aware software: the same builder can access device sensors such as GPS, Bluetooth and NFC so prototypes behave like finished apps. The company also said that later this year Gemini will surface more than 450,000 movies and TV shows and identify where to livestream sports, part of a broader push to fold discovery and recommendations into conversational AI.
The immediate effect will be to accelerate prototyping and personal experimentation — google ai studio lets a person go from idea to a runnable phone app in minutes rather than weeks. Without publishing and backend integrations, however, those prototypes will largely stay private sandboxes for now. If Google follows through on the promised Firebase support and family-and-friends publishing, the tool will move quickly from a personal playground into a lightweight distribution channel; until then it is a fast, highly capable way to make functioning Android app prototypes.
