Google has announced new Android development tools, APIs and platform updates at Android Dev Summit. The announcements cover coding agents, AI-powered app experiences, adaptive app development and Android XR, with tools and APIs spanning stable releases, beta and alpha stages.
Matthew McCullough, Vice President of Product Management for Android Developer, said the updates aim to reduce everyday development friction and give developers more flexibility when building Android apps.
Android Studio and Android CLI add agent support
Android Studio Rabbit 2 is now stable, allowing developers to choose coding agents directly within the IDE. Supported options include Claude Agent, Codex, Antigravity and more than 40 agents compatible with the Agent Client Protocol (ACP).
Android CLI has gained capabilities for device testing, UI validation, performance profiling and app quality analysis.
- Device streaming and Compose previews: Firebase-powered Android device streaming lets developers test apps remotely on real devices. Agents can also render Jetpack Compose previews without running an Android Emulator or deploying an APK.
- Performance profiling: Perfetto system traces help developers investigate cold startup latency, UI jank and background wakeups to analyze app performance and memory usage.
- App quality analysis: The Analyze feature checks app bundles and libraries, including AAs and Js, for quality signals related to obfuscation, optimization and shrinking. It uses the same logic as Google Play to identify potential issues before upload.
- Android skills: More than 30 skills covering over 20 areas help agents handle complex tasks, including permission security, Glance widget layouts and Compose performance optimization.
The Android plugin for Google Antigravity has been updated with the latest Android CLI capabilities, including integrations for viewing Compose previews and interacting with real Android devices. Google has also launched official Android plugins for Claude Code and Codex by OpenAI.
New APIs for agentic Android experiences
Google is introducing APIs that let developers expose app functionality to AI agents and build interactive experiences. The Android intelligence system supports prompted and proactive assistance across the Android ecosystem.
The updates include:
- Connected Apps Platform: Developers can prototype a Connected App to share their app’s services with Android users when Gemini acts as the system AI agent. To publish across supported surfaces, they can sign up for the Connected Apps Platform preview. Developers can build integrations using Android MCP or remote MCP, prototype through Gemini developer mode and register through the Google Connected Console preview program when preparing for production. The platform is intended to connect app services with users across Gemini and Search.
- AppFunctions: Now in beta, Android’s unified tool-calling layer lets developers expose app functionality to on-device or system agents through Android MCP.
- A2UI for Jetpack Compose: Available in alpha, Agent-to-UI (A2UI) lets agents return interactive interfaces built from an app’s own components instead of presenting responses as text. Its native Jetpack Compose renderer integrates these interfaces into Android apps.
- ML Kit Agentic API: Available in alpha, the new API extends the ML Kit Prompt API with Gemini Nano 4 capabilities for on-device agents. It automatically executes functions, saves conversation history, manages the orchestration loop and includes automatic guardrails.
Jetpack Compose and adaptive app development
Google has announced updates to Jetpack Compose and its adaptive development tools to help developers support different screen sizes and device configurations.
- AppState: Now in alpha, AppState brings Compose’s reactive model to the state layer. Its Compose-first architecture building blocks help developers and AI agents preserve state across device configurations.
- Navigation 3 version 1.2: The stable release adds deep-link support and richer predictive back handling.
- Compose 1.13: The upcoming stable release will add built-in adaptive layout support, with Grid and Flexbox moving to stable. It will also introduce Progressive Blur and Formatted Text.
Compose adaptive skill adds Googlebook and foldable support
Google has updated its Compose adaptive skill to help agents adapt existing apps for Googlebook, phones, foldables, tablets, parked cars and XR devices. The update adds desktop-specific improvements and foldable posture optimizations.
The skill can now:
- Audit Compose UI and implement keyboard shortcuts, custom pointer cursors, hover states and tooltips for buttons and controls.
- Handle navigation and migrate layouts to Grid, Flexbox and mediaQuery.
- Manage state across the app.
- Adapt media and camera screens for foldable postures, including tabletop layouts.
Google Play will pay closer attention to how apps and games run on foldable devices next year, providing more opportunities for developers to showcase adaptive implementations. Google recommends that app developers use the Compose adaptive skill and follow the Android adaptive quality guidelines, while game developers should follow the adaptive games guidelines.
McCullough said Google will announce the Google Play Store requirements later this year.
Android XR tools and libraries reach stable
Google and XREAL have launched XREAL AURA, the world’s first wired XR glasses powered by Android XR, alongside updates to Android XR libraries and development tools.
- Core Jetpack XR SDK libraries: Jetpack SceneCore, ARCore for Jetpack XR and XR Runtime now provide production-ready APIs, long-term API stability and backward compatibility.
- Android XR emulators: Stable in Android Studio, the emulators let developers build, iterate and visually validate experiences for XR glasses, audio glasses and display glasses without physical hardware.
- Android XR Engine Hub: The stable hub provides dedicated workflows and access to performance traces across Unity, Godot and Unreal.
- Geospatial APIs: Available in ARCore for Jetpack XR SDK and Unity, these APIs let developers anchor digital content to the real world, with accuracy and local tracking across wired XR glasses and intelligent eyewear.
McCullough said Google welcomes developer feedback on the new tools and capabilities as it continues improving its Android development tools.