FeedworthyAI
Back to feed
Android Developers - Mediumblog

Now in Android #120

Wednesday, August 6, 2025Daniel GalpinView original
agentic aiFeaturednow-in-androidandroidcanarygoogle-play

Canary Channel, Agentic AI in Android Studio, 16 KB Pages, Policy Changes, Compose Tips, Play Games

Welcome to Now in Android, your ongoing guide to what’s new and notable in the world of Android development.

This edition features the Android platform Canary Channel, the first release of Agentic AI in Android Studio, Play updates including 16 KB Page Size requirements and other policy changes, new capabilities around one-time purchases, Play Games Services v2, and Google Play Games on PC, I/O recaps on AI and dev productivity, next gen Samsung devices running Android 16, Compose tips, and more.

Most of the content of this post is available in the form of a video or podcast, so feel free to watch or listen rather than read on. (Or do all three to help you remember! There won’t be a quiz.)

Evolving Android’s early-access programs: Introducing the Canary channel 🐦

Android has introduced a new Canary release channel for its early-access programs, replacing the previous developer preview model. This channel provides earlier and more consistent access to in-development platform features throughout the year.

The new Canary channel offers continuous, rolling over-the-air (OTA) updates for supported Pixel devices, addressing previous limitations where previews required manual flashing and the preview program ended once a platform version reached Beta, requiring manual flashing to enter it again for the next cycle.

You can use the Canary channel to:

  • Access new features and potential behavior changes in their earliest stages.
  • Integrate builds into your Continuous Integration (CI) systems to identify app compatibility issues early.
  • Download Canary SDKs through the SDK Manager and access builds via the Android Studio Canary channel for development and testing.

The Canary channel operates in parallel with the existing Beta program. Canary builds are bleeding-edge and may contain bugs or breaking changes, making them best suited for exploring and trying out apps, rather than primary device use. The Beta channel remains as the place to focus your testing to make sure you’re ready for the next platform launch.

You can get started by flashing your supported Pixel device using the Android Flash Tool or configuring the Android Emulator with a Canary image. You can provide feedback and bug reports through the Google Issue Tracker.

Evolving Android's early-access programs: Introducing the Canary channel

Agentic AI takes Gemini in Android Studio to the next level 🤖

The latest Android Studio Narwhal Feature Drop Canary release includes “Agent Mode,” a new capability that enables the AI to handle complex development tasks. This feature allows you to describe a goal in natural language, and the agent formulates an execution plan that can span multiple project files, utilizing IDE tools for code changes, building, and searching.

You interact with it by describing tasks and then reviewing, accepting, or rejecting the proposed changes, with an option for auto-approve.

Additionally, you can enhance Agent Mode by adding your own Gemini API key, which expands the context window to 1 million tokens with Gemini 2.5 Pro. This expanded context allows Gemini to process more instructions and code. Business tier subscribers automatically receive this benefit through their Gemini Code Assist license.

Agent Mode also supports the Model Context Protocol (MCP), providing a standardized way to integrate with external tools. In this initial release, MCP supports stdio transport for tool interaction. You configure MCP servers via an mcp.json file in your Studio configuration directory.

Agentic AI takes Gemini in Android Studio to the next level

Transition to using 16 KB page sizes for Android apps and games using Android Studio 🚀

Android is transitioning to 16 KB memory page sizes to enhance app performance. Beginning November 1, 2025, new and updated apps developed against the NDK and targeting Android 15+ submitted to Google Play must support 16 KB page sizes. Apps developed exclusively with Kotlin or Java are already compatible.

Devices configured for 16 KB page sizes may see overall performance improvements of 5–10%, including faster app launch times, reduced battery consumption, and quicker camera starts. Since 4KB pages are also aligned to 16KB boundaries, a single application binary can run on both 4 KB and 16 KB devices once updated.

Android Studio provides tools to assist with this transition:

Transition to using 16 KB page sizes for Android apps and games using Android Studio

New tools to help drive success for one-time products ⚙️

Google Play is updating how you manage one-time products, introducing a new three-level hierarchy for defining and selling digital items.

The new model consists of:

  • One-time product: Defines the item being purchased (e.g., “Diamond sword”).
  • Purchase option: Dictates how the item is granted, its price, and availability. You can now use “buy” or “rent” options.
  • Offer: Modifies a purchase option, enabling features like discounts or pre-orders.

This allows you to sell the same product in multiple ways. New capabilities include configuring rental periods, setting up pre-orders with defined dates and discounts, and managing regional pricing and availability for individual purchase options and offers. The concept of a default product price has been removed.

To utilize these new features, you must upgrade to Play Billing Library 8.0 and use the new monetization.onetimeproducts service of the Play Developer API or the updated Play Developer Console. You also need to integrate with the queryProductDetailsAsync API. Existing SKUs can be migrated, but new features are not supported by the older querySkuDetailsAsync API or inappproducts service. Products created via the Play Console UI will be normalized within the new system. Pre-order functionality is currently available through an Early Access Program.

New tools to help drive success for one-time products

Start building for the next generation of Samsung devices 📱

Samsung unveiled the Galaxy Z Fold7, Z Flip7, and Galaxy Watch8 series, offering you new opportunities to build for adaptive experiences. The foldables run Android 16 with an enhanced desktop windowing experience, and you can create adaptive apps for them with Compose Adaptive Layouts, Jetpack Navigation 3, and Window Size Classes. The Galaxy Watch 8 series is the first device to feature Wear OS 6.

Start building for the next generation of Samsung Galaxy devices

Level up your game: Google Play’s Indie Games Fund in Latin America returns for its 4th year 🎮💰

Google Play’s Indie Games Fund in Latin America returns for its fourth year, committing another $2 million to support ten indie game studios in the region. This new round of funding will bring the total investment in Latin American indie games through this program to $8 million USD.

Selected studios will receive between $150,000 and $200,000, along with support from the Google Play team. The program is open to indie game developers who have already launched a game on any platform (mobile, PC, or console).

Applications closed on July 31, 2025, with priority given to submissions received by July 15, 2025.

Level up your game: Google Play's Indie Games Fund in Latin America returns for its 4th year

Upcoming changes to Wear OS watch faces 🔄

Wear OS watch face development is changing to exclusively support the Watch Face Format. As of January 27, 2025, you cannot publish new AndroidX or Wearable Support Library (WSL) watch faces to the Play Store. By January 14, 2026, existing legacy watch faces will no longer be installable from the Play Store, nor will you be able to update them or continue monetizing them. To continue offering your watch faces, you must migrate them to the Watch Face Format and resubmit them to the Play Store by January 14, 2026. Watch Face Studio users should resubmit using version 1.8.7 or newer. The blog post lists resources that are available to assist with migration.

Upcoming changes to Wear OS watch faces

Articles 📚

#WeArePlay: With over 3 billion downloads, meet the people behind Amanotes 🎶

Amanotes, a mobile game company founded by Bill and Silver, has surpassed 3 billion downloads for its rhythm-based music games like Magic Tiles 3. The company aims to make musical performance accessible by combining mobile gaming with rhythm-based taps. Amanotes leveraged tools such as Firebase for A/B testing and Play Console for analytics and custom pricing for its global growth on Google Play.

#WeArePlay: With over 3 billion downloads, meet the people behind Amanotes

Top announcements to know from Google Play at I/O ’25 📈

At I/O ’25, Google Play introduced several updates to help you streamline subscriptions and maximize revenue.

  1. Subscriptions with add-ons enable multi-product checkout, allowing you to sell subscription add-ons with base subscriptions in a single transaction and aligned payment schedule. You can learn more about creating add-ons here.
  2. Showcasing benefits in more places across Play, including the Subscriptions Center, reminder emails, and during purchase and cancellation processes, has reduced voluntary churn by 2%. You can enter your subscription benefit details in Play Console.
  3. New grace period and account hold durations aim to decrease involuntary churn. You can now customize a grace period of up to 30 days and an account hold period of up to 60 days, with the total recovery period not exceeding 60 days. Developers extending their decline recovery period have seen an average 10% reduction in involuntary churn for renewals. You can adjust these settings in Play Console.

Additional initiatives include prompting users to set up payment methods, expanding payment options and market availability, optimizing in-app transactions with AI, and using cart abandonment reminders.

Top announcements to know from Google Play at I/O '25

Get ready for the next generation of gameplay powered by Play Games Services 🚀

Play Games Services (PGS) is updating its features to enhance game discovery and player retention on Google Play. You must migrate from PGS v1 to PGS v2 by May 2026, as games using the v1 SDK will no longer be publishable or updatable after this date, and v1 APIs will be fully deprecated by Q3 2028.

Key updates and features include:

  • Achievements: Now deeply integrated across the Play Store, achievements are discoverable via search and game detail pages, and can drive engagement with Play Points. You should configure at least 15 incremental achievements, with 5 achievable within the first two hours of gameplay. Bulk configuration is available in Play Console.
  • Promotional Content: Audience targeting capabilities have been enhanced. You can use the Play Grouping API to create custom segments based on gameplay context for tailored re-engagement promotions.
  • Sign-in: PGS offers seamless sign-in and progress syncing across Android devices. The Recall API helps you manage links between PGS and your in-game accounts.

Future developments include AI-powered avatars that you can integrate into player profiles within your game.

Get ready for the next generation of gameplay powered by Play Games Services

How Mecha BREAK is driving PC-only growth on Google Play Games 🤖

Amazing Seasun Games is launching its action shooting game, Mecha BREAK, on Google Play Games on PC as part of its multiplatform PC and console release.

The game uses Google Play Games on PC for its launch strategy, including pre-registration campaigns that attracted over 3.5 million players during beta. You can access the native PC program to receive up to 15% in additional earnback, along with support for PC game development, distribution, and growth. This support includes managing PC builds, releases, and store listings through Play Console, and accessing PC-specific sales reports.

The platform provides a secure infrastructure with tools like Play Integrity API, offering malware protection and anti-cheat capabilities. Mecha BREAK itself delivers a high-fidelity PC experience with a proprietary 3D engine, offering diverse combat styles such as 6v6 hero battles, 3v3 matches, and a PvPvE “Mashmak” extraction mode.

How Mecha BREAK is driving PC-only growth on Google Play Games

Top 3 Updates for Android Developer Productivity @ Google I/O ’25 🤖

Here are the top 3 updates for Android developer productivity from Google I/O 2025:

  1. Gemini in Android Studio: Gemini-powered AI is now available to simplify building and testing code, with features like “Journeys for Android Studio,” a “Version Upgrade Agent,” and “Agent Mode” for complex tasks. It can also generate Jetpack Compose previews and transform UI code using natural language.
  2. New Jetpack Compose release: The latest stable version includes Navigation 3 for more flexibility, a Material 3 Expressive update to enhance user experience, and new features like autofill support and auto-sizing text, along with performance improvements.
  3. Kotlin Multiplatform (KMP) shared module template: A new template in Android Studio helps developers share business logic between Android and iOS apps, reducing development time. This is supported by updated Jetpack libraries and new codelabs.

Top 3 updates for Android developer productivity at Google I/O '25

Top 3 things to know for AI on Android at Google I/O ’25 🤖

Google I/O ’25 announced three key updates for AI on Android development.

  1. Gemini Nano for on-device AI experiences: New ML Kit GenAI APIs powered by Gemini Nano are available for tasks like text summarization, proofreading, and image description directly on devices.
  2. Google AI Edge for custom models: The Google AI Edge platform now offers improved support for on-device hardware accelerators and a new AI Edge Portal for benchmarking. You can also manage custom model downloads efficiently with Play for On-Device AI, now in beta.
  3. Firebase AI Logic for cloud AI: For advanced generative AI use cases, Firebase AI Logic provides a simplified, secure way to integrate larger models like Gemini Flash, Gemini Pro, and Imagen, including support for conversational AI with Gemini Live API and visual asset generation with Imagen.

Additionally, the open-source Androidify app demonstrates building AI-driven experiences with these new APIs, and more examples are available in the Android AI Sample Catalog.

Top 3 things to know for AI on Android at Google I/O '25

Videos 📹

Enrich your app with live updates and widgets

Ash and Aaron cover the new capabilities for Widgets and Live Updates.

Widgets now offer Canonical Layouts for faster development and Generated Previews (enhanced in Glance 1.2 for Android 15+) that display real user data, requiring careful management of memory and rate limits.

For Live Updates & Notifications, a new ProgressStyleTemplate streamlines progress indicators for time-sensitive, user-initiated journeys (e.g., delivery, ride-share) — not for ads or spam. Developers can programmatically determine if notifications meet Live Update criteria and customize a new prominent status chip for quick user access.

Reaching billions: Multiplatform strategies and performance tools for Android

Google Play now enables multi-platform Android game distribution from a single Android App Bundle, reaching billions of users across phones, tablets, foldables, and PCs. Google Play Games on PC is nearing general availability, offering automatic keyboard-to-touch translation for existing games and optional native mouse support (via a manifest flag) for an enhanced PC experience.

Navigating the future with Jetpack Navigation 3 — Core Concepts

Jetpack Navigation 3 is now in Alpha, offering a new state-based approach specifically designed to align with Jetpack Compose’s declarative nature.

Key newsworthy points for Android app developers:

  • Decoupled Logic: It provides a strong separation between navigation logic and UI, improving testability and flexibility.
  • State as Single Source of Truth: Advocates for your app’s state (including navigation) as the single source of truth, moving away from event-based navigation issues.
  • Developer-Owned Backstack: Developers now own the backstack state, giving more control and simplifying complex scenarios like type-safety across modules and multi-pane layouts.
  • Flexible Building Blocks: Provides building blocks that integrate better with Compose’s patterns, allowing a more robust and developer-controlled navigation solution.

Developers are encouraged to explore the Alpha and provide feedback.

Compose Preview | Jetpack Compose Tips

Compose Previews in Android Studio significantly accelerate UI development by allowing you to test Jetpack Compose UI changes instantly without full builds.

Key features include:

Crucial best practices include:

  • Avoiding direct ViewModel instantiation in previews.
  • Using @PreviewParameter for efficient data provision.
  • Creating a “sticker sheet” with multiple @Preview annotations for visual component documentation and faster team onboarding.

Embedded Layout Inspector | Jetpack Compose Tips

The Embedded Layout Inspector in Android Studio is a crucial tool for debugging Jetpack Compose UIs, allowing you to inspect layout hierarchy. It’s accessed via the Running Devices Tool Window while your app is running.

Build a Runtime-Enabled SDK

Android 14 introduced SDK Runtime, a new technology enhancing privacy and security by isolating third-party SDKs (called Runtime-enabled SDKs or RE SDKs) in a secure sandbox. Apps/SDKs (called Runtime-aware SDKs or RA SDKs) interact with these RE SDKs via automatically generated “shims,” with backward compatibility provided by Jetpack.

For SDK developers, building an RE SDK requires significant changes: a specific project structure using Android SDK Bundles (ASB), new Gradle configurations (including Privacy Sandbox plugins and specialized Jetpack dependencies), and declaring APIs with new annotations (@PrivacySandboxService, @PrivacySandboxInterface). Special considerations apply for UI and launching activities, and a specific entry point must be implemented. This marks a fundamental shift in how SDKs will be built and integrated on Android.

Google Play PolicyBytes — July 2025 policy updates

Key Takeaways:

  • Body Sensor Permissions are Evolving: Starting with Android 16, Google is replacing broad body sensor permissions with new, specific, privacy-focused ones (e.g., android.permission.health.READ_HEART_RATE). Apps accessing sensitive health data (heart rate, oxygen, skin temp) will need to adapt to these new permissions, requiring approved, user-benefiting use cases.
  • Blockchain/Crypto Apps — Country-Specific Rules: A new Help Center article provides country-specific requirements for apps offering blockchain content (crypto exchanges/wallets). Developers in this space should be aware of potential location-specific forms.
  • AI-Generated Content Best Practices: New guidance is available on best practices for apps that include AI-generated content, focusing on protecting users and apps.
  • Content Ratings & Ads: Content ratings only reflect your app’s internal content, not ads. Ensure ad content displayed within your app is not significantly more mature than your app’s content rating.
  • Upcoming Deadlines: Remember to review 2025 deadlines for various policies, including Photo/Video Permissions, Lines of Credit, News/Magazines, and Health Content & Services.

Policy announcement: July 10, 2025

Attention Personal Loan developers targeting Pakistan!

Google Play has updated its policy for personal loan apps targeting Pakistan: Apps promoting short-term personal loans (repayment within 60 days or less) are now generally prohibited. If you have such apps, make sure to review your compliance.

Blockchain-based content policy help center

For Android apps dealing with blockchain content (like crypto exchanges or software wallets), Google Play now provides a Blockchain-based Content Policy Help Center. This resource details country-specific compliance requirements, including potential location-specific forms. If your app targets a location not explicitly listed, you must declare it as a crypto exchange or software wallet in the Financial Features declaration. See goo.gle/blockchain-policy for full details.

Broadening audiences through expansions (Playtime 2024)

Facing declining mobile gaming growth in 2023, Google Play highlights ‘genre expansion’ as a crucial strategy for developers. This involves integrating gameplay, monetization, or aesthetics across genres to acquire broader audiences, deepen engagement, and boost monetization (e.g., 4X games adding casual mini-games). You should focus on early UA testing, understanding player motivations, ensuring seamless integration, and utilizing Play Console tools like store listing experiments and staged rollouts for effective iteration.

Unlock the full potential of your game with Vulkan and ADPF

Google’s Android team highlights Vulkan and the Adaptive Performance Framework (ADPF) as crucial for modern Android game development. Vulkan is an advanced graphics API offering increased efficiency and access to advanced GPU features, enabling you to maximize your games’ potential on contemporary hardware.

AndroidX Releases 🚀

Here’s a summary of the recent AndroidX changes:

Critical Bug Fixes & Stability Improvements:

Paging (3.4.0-alpha01): Several critical bugs affecting UI stability and data loading have been fixed:

  1. Resolved IndexOutOfBoundsException when scrolling while updating RecyclerView.
  2. Fixed issues preventing more loads from triggering when refreshing while scrolling.
  3. Addressed crashes occurring when scrolling while refreshing Paging items.

Room (2.8.0-alpha01): Fixed a deadlock that could occasionally occur when using suspending transactions with the AndroidSQLiteDriver.

Significant New Features & API Enhancements:

WebKit (1.15.0-alpha01): Major enhancements for WebView control:

  1. HTTP Header Control: A new API (Profile.setOriginMatchedHeader) to set HTTP header name-value pairs on requests to specific origins within WebViews and service workers.
  2. Improved Cookie Handling: Developers can now include Cookie headers in WebResourceRequest.getRequestHeaders() and process Set-Cookie values from WebResourceResponseCompat.setCookies within shouldInterceptRequest callbacks. This greatly improves control over cookie management in intercepted WebView requests.
  3. Renderer Warm-up: An experimental API to warm up the WebView renderer process for a Profile, improving performance for complex WebView lifecycles.

Room (2.8.0-alpha01): Introduced androidx.room:room-sqlite-wrapper and the RoomDatabase.getSupportWrapper() extension function. This allows developers adopting SQLiteDriver to still obtain a SupportSQLiteDatabase wrapper, easing incremental migration for codebases that have extensive SupportSQLite API usages.

Expanded Kotlin Multiplatform (KMP) Support (Medium Priority, High for KMP Developers):

Wear OS Specific Updates (Medium Priority, High for Wear Developers):

Other Notable Changes (Lower Priority for general app development, but important for specific use cases):

  • SQLite (2.6.0-alpha01): Adds an API (SQLiteConnection.hasActiveTransaction()) to check if a SQLiteConnection has an active transaction, which can aid in managing complex transaction logic.
  • Test UI Automator (1.0.0-alpha01): A new uiautomator-shell library allows executing shell commands as a shell user, backporting UiAutomation#executeShellCommandRwe (from API 34) for testing and instrumentation purposes.

Android Developers Backstage

The Android Developers Backstage podcast is back with two more episodes.

Bot to the future: Two years later

AI-powered Gemini features in Android Studio have significantly evolved, offering comprehensive developer support from design to maintenance. Key advancements include intelligent code completion, design tool integrations (e.g., composable previews), and unit test scenario generation.

Newsworthy features:

  • Version Upgrade Assistant: Automates dependency updates and fixes compilation errors, reducing developer “toil.”
  • Agent Mode (public preview): Allows Gemini to directly modify code within the IDE based on instructions, with developer approval required for changes.
  • Privacy: Gemini operates within the IDE, keeping code private without external transmission for model training, with enhanced privacy options for businesses.

Back it up, bring it back: No data left behind

Android Studio now automates app data backup and restore testing, eliminating cumbersome manual ADB commands. This significantly simplifies validating your app’s data persistence (settings, game states, permissions) across devices and emulators. Developers can configure backup scope via the manifest, differentiating between cloud (25MB) and device-to-device (1–2GB) transfers. Customization of backup descriptors is possible for complex scenarios, though sensitive data is excluded for security. This new tooling streamlines testing for reliable data transfer and a seamless user experience.

Now then… 👋

That’s it for this edition of Now in Android, including the Android platform Canary Channel, the first release of Agentic AI in Android Studio, Play updates including 16 KB Page Size requirements and other policy changes, new capabilities around one-time purchases, Play Games Services v2, and Google Play Games on PC, I/O recaps on AI and dev productivity, next gen Samsung devices running Android 16, and Compose tips.

Check back soon for your next update from the Android developer universe!

Now in Android #120 was originally published in Android Developers on Medium, where people are continuing the conversation by highlighting and responding to this story.