Android Studio, an integrated development environment (IDE) for Android app developers, has obtained an AI upgrade. Google announced it at Google I/O 2025, a developer meeting on Tuesday.
In addition to deploying the latest Gemini 2.5 Pro models, Android Studio acquires a new “agent AI” feature called Journeys, and quickly introduces “agent mode” for more complex development tasks.
With Gemini, Journeys allows developers to test their apps by explaining actions and assertions in natural language for the user's journey across the app. Gemini then runs the test.
Image credit: Google
This feature, Google said, makes it easier for developers to test their apps without having to write extensive code. The company warns that this is still experimental, but ultimately it's about increasing the speed at which high quality code is shipped, while reducing the time it takes to test, verify or reproduce the issue.
The tests can be run on physical or virtual Android devices, and the results will be displayed directly in the IDE, Google says.
Image credit: Google
Another immediate arrival includes an autonomous AI feature with Gemini, called Agent Mode. This allows developers to use a variety of tools to handle more complex multi-stage development tasks. For example, if a developer is trying to integrate a new API, the agent can come up with an execution plan that will add the required dependencies, edit the files, and fix the bug.
Other AI features that appear in Android Studio include Gemini-driven improvements to the “Crash Insights” feature in the APP's Quality Insights panel. This allows AI to determine what the source code of an app crashed and suggest fixes.
Image credit: Google
Additionally, Google now allows developers to try out experimental AI features using the new “Studio Lab” menu in the Android Studio settings menu. This option is only available for stable releases starting with the release codename NARWHAL.
Image credit: Google
Another experiment currently available is the public preview of Android Studio Cloud. Accessed through Firebase Studio, the new service streams Linux machines running Android Studio to a web browser, allowing Android development wherever you have access to your internet connection.
The version of the upgrade agent will soon arrive as part of Android Studio's Gemini, which will help you automate dependency upgrades.
Gemini automatically generates preview code for JetPack composing, converts UI code within the preview environment using natural language, attaches image files (such as UI mockups and screenshots) to AI prompts, attach project files as the context for chatting with Gemini, and sets preferred coding style or output format with Gemini's new “rules” feature.
Image credit: Google
The company also deploys an enterprise-ready version of Android Studio powered by AI, and is launching Gemini in Android Studio for Business. This allows teams to deploy teams while keeping their data safe when subscribing to the standard or enterprise edition.
Other updates include relocatable previews of the configuration for preview and navigation improvements, an embedded Android XR emulator that launches by default in embedded state, and upgrades backup and restore, backup and sync, among other things. The Kotlin Multiplatform on Android also has a small number of improvements.
Furthermore, Google says it will help developers prepare for Android's 16KB page size (changes to Android's underlying architecture). Use early warnings and tools to test your app in a new environment.