Android Studio Rabbit 2 Launches With Bring Your Own Agent, XR Emulators, and Advanced Code Review Tools

Google has officially released Android Studio Rabbit 2, marking a major milestone for the popular integrated development environment as it transitions from preview status into a stable release ready for full-scale production use. Announced alongside the live-streamed sessions of the Android Dev Summit, this latest version introduces a comprehensive suite of powerful capabilities designed to streamline modern mobile and spatial application development. Centered heavily around agentic workflows, hardware virtualization for mixed reality, and enhanced code governance, Rabbit 2 aims to address the evolving complexities faced by professional engineering teams worldwide.

The release arrives at a time when developers are increasingly relying on artificial intelligence and specialized coding assistants to accelerate development cycles. By focusing on deep integration with the developer environment, the engineering team behind Android Studio has sought to provide tools that not only speed up code generation but also maintain high standards of quality, security, and performance.

Bring Your Own Agent (BYOA): Integrate Your Preferred Coding Agents

One of the most notable updates in Android Studio Rabbit 2 is the introduction of Bring Your Own Agent, commonly referred to as BYOA. Recognizing that different teams and projects require different levels of flexibility, cost efficiency, and governance, Google has designed the new environment to support a vast array of Agent Client Protocol (ACP) compliant coding assistants.

Developers are no longer locked into a single ecosystem. Instead, they can choose from over forty different coding agents directly within Android Studio. This roster includes prominent options such as Claude Agent from Anthropic, Codex from OpenAI, Junie, and Google’s own Antigravity. Whether a development shop requires a specific agent to align with stringent enterprise agreements or simply wants to leverage the cost efficiency of a particular model for an independent project, the infrastructure is now in place to support it.

These integrated agents are given deep access to the core components of Android Studio. They can interact directly with the Android Emulator, underlying build systems, and Logcat to diagnose issues in real time. Furthermore, they are equipped to utilize specialized Android skills to execute complex, multi-step engineering tasks, blending the raw predictive power of large language models with the deterministic utility of native development tools. Getting started with the feature requires only a brief authentication step within the agent configuration menu, opening the door to a highly personalized AI-assisted workflow.

Build for Android XR with Emulators

As the computing landscape continues to expand beyond traditional mobile screens into spatial environments, Android Studio Rabbit 2 brings robust stability to its Android XR Emulators. Developers can now build, iterate, and rigorously test immersive experiences for XR glasses, audio glasses, and advanced display glasses directly from their laptops without requiring physical hardware on hand at all times.

The updated display glasses emulator introduces powerful customization features, allowing engineers to test their applications against realistic physical environments from anywhere in the world. Users can map their virtual surroundings using custom 360-degree photos or feed live video streams from a connected camera directly into the emulator environment. To ensure applications adhere to modern privacy and user experience standards, Google has also incorporated hardware-level LED indicators that visually represent active microphone and camera states within the virtual space.

Additionally, the underlying pairing wizard has been significantly overhauled to allow virtual phones to connect with virtual audio and display glasses more quickly and reliably. These enhancements are intended to lower the barrier of entry for mobile developers looking to transition their existing skill sets into the rapidly growing field of spatial computing.

Discover, Install, and Manage Android Skills in Android Studio

To make agentic workflows more practical and tailored to specific mobile engineering challenges, Android Studio Rabbit 2 introduces a dedicated skills manager. This interface allows developers to discover, install, and manage specialized procedural workflows and guidance packages that direct how coding agents operate within the workspace.

The IDE ships pre-loaded with a variety of native Android skills and Firebase agent skills designed to handle heavy lifting tasks. These include automating the complex migration of legacy XML Views to Jetpack Compose, upgrading projects to the latest Android Gradle Plugin (AGP) versions, implementing adaptive layouts tailored for foldables and tablets, and integrating complex hardware and cloud features like CameraX and Firebase services.

Developers can invoke these skills instantly during chat interactions by using simple command symbols, or they can manage their installed library through a centralized editor panel. This granular control ensures that the automated assistants are following established best practices rather than generating generic or potentially problematic code.

Agentic Fixes with LeakCanary

Memory leaks have long been one of the most persistent and frustrating performance bottlenecks in Android development. In Rabbit 2, Google has tackled this issue by integrating LeakCanary directly into the built-in development agent within the Android Studio Profiler.

Instead of merely pointing out that a memory leak has occurred and leaving the developer to manually trace object references and stack traces, the updated Profiler leverages artificial intelligence to analyze the heap dump. LeakCanary can now provide clear, narrative explanations detailing the root causes of the leak and suggest specific, context-aware code modifications to resolve the issue. By bringing these diagnostic and corrective workflows directly into the IDE, teams can catch and eliminate memory regressions much earlier in the development lifecycle.

Gemini Enterprise in Android Studio

For larger engineering organizations that demand strict data privacy controls, comprehensive governance frameworks, and integration with proprietary internal knowledge bases, Android Studio continues to provide robust support for Gemini Enterprise. Originally introduced in earlier iterations of the platform, this tier allows enterprise teams to securely connect their development environments to internal design systems, proprietary SDKs, and private repositories.

This deep contextual awareness ensures that code suggestions, automated reviews, and architectural guidance remain tightly aligned with company-specific standards rather than relying solely on public domain knowledge.

Go from Raw Diffs to Guided Code Reviews with Change Review

While generative artificial intelligence and autonomous coding agents have made writing and refactoring code faster than ever, they have inadvertently created a new bottleneck for engineering teams: the review and verification process. Reading through thousands of lines of raw code diffs generated by various agents or team members can be an exhausting and error-prone endeavor.

To solve this challenge, Android Studio Rabbit 2 introduces AI Change Review, a tool designed to streamline the verification process regardless of where code changes originate in the workspace. Developers can launch a review from multiple visual entry points or directly from the agent chat window.

The system initiates the process by surfacing a curated list of recommended fixes, where the agent independently reviews its own output, pending commits, or recent changes under review. Once these initial suggestions have been addressed, developers can trigger a guided walkthrough. Rather than forcing reviewers to examine modified files in alphabetical or chronological order, the guided walkthrough intelligently groups and prioritizes related diff chunks across the entire project into logical, themed review units. This structured approach significantly reduces cognitive load and helps teams maintain rigorous code quality standards without sacrificing the velocity gained through modern automation tools.

Getting Started

The release of Android Studio Rabbit 2 is currently rolling out globally. Developers interested in adopting the new production-ready features can download the latest stable build directly through the official Android developer portal. Google encourages developers encountering any bugs or unexpected behaviors to submit detailed reports through the standard issue-tracking channels.

As the developer community begins adopting these new tools, further discussions, technical deep dives, and live demonstrations are being shared across official channels, including the Android Dev Summit live streams and community platforms on LinkedIn, YouTube, and X. With Rabbit 2 now established in the stable channel, the mobile development ecosystem enters a new phase characterized by deeper automation, spatial readiness, and more cohesive engineering workflows.