Jetpack Navigation 3 is a new Google navigation library that is fundamentally different from previous versions. The main idea of Nav3 is simple: you have a NavBackStack — a regular mutable list where each element represents a screen in your application.Jetpack Navigation 3 is a new Google navigation library that is fundamentally different from previous versions. The main idea of Nav3 is simple: you have a NavBackStack — a regular mutable list where each element represents a screen in your application.

Nav3 Router: Convenient Navigation on Top of Jetpack Navigation 3

2025/09/17 20:00

What is Jetpack Navigation 3?

Jetpack Navigation 3 is a new Google navigation library that is fundamentally different from previous versions. The main idea of Nav3 is simple: you have a NavBackStack — a regular mutable list where each element represents a screen in your application.

\ You add and remove elements from this list, and the UI automatically updates. Each screen is represented as a NavKey — a regular Kotlin class.

\ This gives you full control over navigation, but requires writing quite a lot of boilerplate code for typical operations.

Why Working Directly with NavBackStack is Inconvenient

Let's look at what the code looks like when working directly with NavBackStack:

@Composable fun MyApp() {     val backStack = rememberNavBackStack(Screen.Home)      // Add a screen     backStack.add(Screen.Details("123"))      // Go back     backStack.removeLastOrNull()      // Replace current screen     backStack.set(backStack.lastIndex, Screen.Success) } 

Problems begin when you need to trigger navigation from a ViewModel. You'll have to either pass NavBackStack to the ViewModel (which, in my understanding, violates architectural principles, as I believe the ViewModel shouldn't know about Compose-specific things), or create intermediate callbacks for each navigation action.

\ Additionally, when working with the stack directly, it's easy to forget to handle edge cases.

How Nav3 Router Simplifies the Work

Nav3 Router is a thin wrapper over Navigation 3 that provides a familiar API for navigation. Instead of thinking about indices and list operations, you simply say "go to screen X" or "go back."

\ Important point: Nav3 Router doesn't create its own stack. It works with the same NavBackStack that Navigation 3 provides, just making it more convenient to work with. When you call router.push(Screen.Details), the library translates this into the corresponding operation with the original stack.

\ Main advantages:

  • Can be used from ViewModel
  • Navigation commands are buffered if UI is temporarily unavailable (for example, during screen rotation)
  • All stack operations happen atomically
  • Clear API
  • Flexibility in modification and adding custom behavior

Installation

Nav3 Router is available on Maven Central. Add the dependency to your build.gradle.kts:

// For shared module in KMP project kotlin {     sourceSets {         commonMain.dependencies {             implementation("io.github.arttttt.nav3router:nav3router:1.0.0")         }     } }  // For Android-only project dependencies {     implementation("io.github.arttttt.nav3router:nav3router:1.0.0") } 

\ The library source code is available on GitHub: github.com/arttttt/Nav3Router

How Nav3 Router is Structured

The library consists of three main parts, each solving its own task:

Router — Developer Interface

Router provides methods like push(), pop(), replace(). When you call these methods, Router creates corresponding commands and sends them down the chain. Router itself knows nothing about how navigation will be executed — this allows using it from anywhere.

CommandQueue — Buffer Between Commands and Their Execution

CommandQueue solves the timing problem. Imagine: the user pressed a button during screen rotation. The UI is being recreated, and the navigator is temporarily unavailable. CommandQueue will save the command and execute it as soon as the navigator is ready again. Without this, the command would simply be lost.

// Simplified queue logic class CommandQueue<T : Any> {     private var navigator: Navigator<T>? = null     private val pending = mutableListOf<Command<T>>()      fun executeCommand(command: Command<T>) {         if (navigator != null) {             navigator.apply(command)  // Navigator exists - execute immediately         } else {             pending.add(command)       // No - save for later         }     } } 

Navigator takes commands and applies them to NavBackStack. Important detail: it first creates a copy of the current stack, applies all commands to it, and only then atomically replaces the original stack with the modified copy. This guarantees that the UI will never see intermediate stack states.

// Simplified Navigator logic fun applyCommands(commands: Array<Command>) {     val stackCopy = backStack.toMutableList()  // Work with a copy      for (command in commands) {         when (command) {             is Push -> stackCopy.add(command.screen)             is Pop -> stackCopy.removeLastOrNull()             // ... other commands         }     }      backStack.swap(stackCopy)  // Atomically apply changes } 

Getting Started With Nav3 Router

The simplest way — don't even create Router manually. Nav3Host will do it for you:

@Composable fun App() {     val backStack = rememberNavBackStack(Screen.Home)      // Nav3Host will create Router automatically     Nav3Host(backStack = backStack) { backStack, onBack, router ->         NavDisplay(             backStack = backStack,             onBack = onBack,             entryProvider = entryProvider {                 entry<Screen.Home> {                     HomeScreen(                         onOpenDetails = {                              router.push(Screen.Details)  // Use router                         }                     )                 }                  entry<Screen.Details> {                     DetailsScreen(                         onBack = { router.pop() }                     )                 }             }         )     } } 

For more complex applications, it makes sense to create a Router through DI and pass it to the ViewModel:

\ Define screens

@Serializable sealed interface Screen : NavKey {     @Serializable     data object Home : Screen      @Serializable     data class Product(val id: String) : Screen      @Serializable     data object Cart : Screen } 

\ Pass router to Nav3Host.

@Composable fun App() {     val backStack = rememberNavBackStack(Screen.Home)     val router: Router<Screen> = getSomehowUsingDI()      // Pass Router to Nav3Host     Nav3Host(         backStack = backStack,         router = router,     ) { backStack, onBack, _ ->         NavDisplay(             backStack = backStack,             onBack = onBack,             entryProvider = entryProvider {                 entry<Screen.Home> { HomeScreen() }                 entry<Screen.Details> { DetailsScreen() }             }         )     } } 

\ ViewModel receives Router through constructor.

class ProductViewModel(     private val router: Router<Screen>,     private val cartRepository: CartRepository ) : ViewModel() {      fun addToCart(productId: String) {         viewModelScope.launch {             cartRepository.add(productId)             router.push(Screen.Cart)  // Navigation from ViewModel         }     } } 

\ In UI, just use ViewModel.

@Composable fun ProductScreen(viewModel: ProductViewModel = koinViewModel()) {     Button(onClick = { viewModel.addToCart(productId) }) {         Text("Add to Cart")     } } 

Examples of Typical Scenarios

Simple Forward-Back Navigation

// Navigate to a new screen router.push(Screen.Details(productId))  // Go back router.pop()  // Navigate with replacement of current screen (can't go back) router.replaceCurrent(Screen.Success) 

Working with Screen Chains

// Open multiple screens at once router.push(     Screen.Category("electronics"),     Screen.Product("laptop-123"),     Screen.Reviews("laptop-123") )  // Return to a specific screen // Will remove all screens above Product from the stack router.popTo(Screen.Product("laptop-123")) 

Checkout Scenario

@Composable fun CheckoutScreen(router: Router<Screen>) {     Button(         onClick = {             // After checkout we need to:             // 1. Show confirmation screen             // 2. Prevent going back to cart              router.replaceStack(                 Screen.Home,                 Screen.OrderSuccess(orderId)             )             // Now only Home and OrderSuccess are in the stack         }     ) {         Text("Place Order")     } } 

Exiting Nested Navigation

// User is deep in settings: // Home -> Settings -> Account -> Privacy -> DataManagement  // "Done" button should return to home Button(     onClick = {         // Will leave only root (Home)         router.clearStack()     } ) {     Text("Done") }  // Or if you need to close the app from anywhere Button(     onClick = {         // Will leave only current screen and trigger system back         router.dropStack()     } ) {     Text("Exit") } 

Bonus: SceneStrategy and Dialogs

So far, we've only talked about simple navigation between screens. But what if you need to show a dialog or bottom sheet? This is where the SceneStrategy concept from Navigation 3 comes to help.

What is SceneStrategy?

SceneStrategy is a mechanism that determines exactly how screens from your stack will be displayed. By default, Navigation 3 uses SinglePaneSceneStrategy, which simply shows the last screen from the stack. But you can create your own strategies for more complex scenarios.

\ Think of SceneStrategy as a director who looks at your stack of screens and decides: "Okay, these three screens we show normally, but this last one — as a modal window on top of the previous ones". This allows representing different UI patterns with the same stack.

Creating a Strategy for ModalBottomSheet

Let's create a strategy that will show certain screens as bottom sheets. First, let's define how we'll mark such screens:

@Serializable sealed interface Screen : NavKey {     @Serializable     data object Home : Screen      @Serializable     data class Product(val id: String) : Screen      // This screen will be shown as bottom sheet     @Serializable     data object Filters : Screen } 

\ Now, let's create the strategy itself. It will check the metadata of the last screen and, if it finds a special marker, show it as a bottom sheet:

class BottomSheetSceneStrategy<T : Any> : SceneStrategy<T> {      companion object {         // Metadata key by which we identify bottom sheet         private const val BOTTOM_SHEET_KEY = "bottomsheet"          // Helper function to create metadata         fun bottomSheet(): Map<String, Any> {             return mapOf(BOTTOM_SHEET_KEY to true)         }     }      @Composable     override fun calculateScene(         entries: List<NavEntry<T>>,         onBack: (Int) -> Unit     ): Scene<T>? {         val lastEntry = entries.lastOrNull() ?: return null          // Check if the last screen has bottom sheet marker         val isBottomSheet = lastEntry.metadata[BOTTOM_SHEET_KEY] as? Boolean          if (isBottomSheet == true) {             // Return special Scene for bottom sheet             return BottomSheetScene(                 entry = lastEntry,                 previousEntries = entries.dropLast(1),                 onBack = onBack             )         }          // This is not a bottom sheet, let another strategy handle it         return null     } } 

Combining Multiple Strategies

In a real application, you might need bottom sheets, dialogs, and regular screens. For this, you can create a delegate strategy that will choose the right strategy for each screen:

class DelegatedScreenStrategy<T : Any>(     private val strategyMap: Map<String, SceneStrategy<T>>,     private val fallbackStrategy: SceneStrategy<T> ) : SceneStrategy<T> {      @Composable     override fun calculateScene(         entries: List<NavEntry<T>>,         onBack: (Int) -> Unit     ): Scene<T>? {         val lastEntry = entries.lastOrNull() ?: return null          // Check all keys in metadata         for (key in lastEntry.metadata.keys) {             val strategy = strategyMap[key]             if (strategy != null) {                 // Found suitable strategy                 return strategy.calculateScene(entries, onBack)             }         }          // Use default strategy         return fallbackStrategy.calculateScene(entries, onBack)     } } 

Using in Application

Now, let's put it all together. Here's what using bottom sheet looks like in a real application:

@Composable fun ShoppingApp() {     val backStack = rememberNavBackStack(Screen.Home)     val router = rememberRouter<Screen>()      Nav3Host(         backStack = backStack,         router = router     ) { backStack, onBack, router ->         NavDisplay(             backStack = backStack,             onBack = onBack,             // Use our combined strategy             sceneStrategy = DelegatedScreenStrategy(                 strategyMap = mapOf(                     "bottomsheet" to BottomSheetSceneStrategy(),                     "dialog" to DialogSceneStrategy()  // Navigation 3 already has this strategy                 ),                 fallbackStrategy = SinglePaneSceneStrategy()  // Regular screens             ),             entryProvider = entryProvider {                 entry<Screen.Home> {                     HomeScreen(                         onOpenFilters = {                             // Open filters as bottom sheet                             router.push(Screen.Filters)                         }                     )                 }                  entry<Screen.Product> { screen ->                     ProductScreen(productId = screen.id)                 }                  // Specify that Filters should be bottom sheet                 entry<Screen.Filters>(                     metadata = BottomSheetSceneStrategy.bottomSheet()                 ) {                     FiltersContent(                         onApply = { filters ->                             // Apply filters and close bottom sheet                             applyFilters(filters)                             router.pop()                         }                     )                 }             }         )     } } 

\ What's happening here? When you call router.push(Screen.Filters), the screen is added to the stack as usual. But thanks to metadata and our strategy, the UI understands that this screen needs to be shown as a bottom sheet on top of the previous screen, rather than replacing it completely.

\ When calling router.pop() the bottom sheet will close, and you'll return to the previous screen. From Router's point of view, this is regular back navigation, but visually it looks like closing a modal window.

Advantages of This Approach

Using SceneStrategy provides several important advantages. First, your navigation logic remains simple — you still use push and pop without thinking about how exactly the screen will be shown. Second, navigation state remains consistent — bottom sheet is just another screen in the stack that is properly saved during screen rotation or process kill. And finally, it provides great flexibility — you can easily change how a screen is displayed by just changing its metadata, without touching the navigation logic.

\ This approach is especially useful when the same screen can be shown differently depending on context. For example, a login screen can be a regular screen on first app launch and a modal dialog when attempting to perform an action that requires authorization.

Why You Should Use Nav3 Router

Nav3 Router doesn't try to replace Navigation 3 or add new features. Its task is to make working with navigation convenient and predictable. You get a simple API that can be used from any layer of the application, automatic handling of timing issues, and the ability to easily test navigation logic.

\ At the same time, under the hood, regular Navigation 3 still works with all its capabilities: state saving, animation support, and proper handling of the system "Back" button.

\ If you're already using Navigation 3 or planning to migrate to it, Nav3 Router will help make this experience more pleasant without adding unnecessary complexity to the project.

  • GitHub repository: github.com/arttttt/Nav3Router
  • Usage examples: see sample folder in the repository

\

Disclaimer: The articles reposted on this site are sourced from public platforms and are provided for informational purposes only. They do not necessarily reflect the views of MEXC. All rights remain with the original authors. If you believe any content infringes on third-party rights, please contact service@support.mexc.com for removal. MEXC makes no guarantees regarding the accuracy, completeness, or timeliness of the content and is not responsible for any actions taken based on the information provided. The content does not constitute financial, legal, or other professional advice, nor should it be considered a recommendation or endorsement by MEXC.

You May Also Like

Strive CEO Urges MSCI to Reconsider Bitcoin-Holding Firms’ Index Exclusion

Strive CEO Urges MSCI to Reconsider Bitcoin-Holding Firms’ Index Exclusion

The post Strive CEO Urges MSCI to Reconsider Bitcoin-Holding Firms’ Index Exclusion appeared on BitcoinEthereumNews.com. MSCI’s proposed Bitcoin exclusion would bar companies with over 50% digital asset holdings from indexes, potentially costing firms like Strategy $2.8 billion in inflows. Strive CEO Matt Cole urges MSCI to let the market decide, emphasizing Bitcoin holders’ roles in AI infrastructure and structured finance growth. Strive’s letter to MSCI argues exclusion limits passive investors’ access to high-growth sectors like AI and digital finance. Nasdaq-listed Strive, the 14th-largest Bitcoin treasury firm, highlights how miners are diversifying into AI power infrastructure. The 50% threshold is unworkable due to Bitcoin’s volatility, causing index flickering and higher costs; JPMorgan analysts estimate significant losses for affected firms. Discover MSCI Bitcoin exclusion proposal details and Strive’s pushback. Learn impacts on Bitcoin treasury firms and AI diversification. Stay informed on crypto index changes—read now for investment insights. What is the MSCI Bitcoin Exclusion Proposal? The MSCI Bitcoin exclusion proposal seeks to exclude companies from its indexes if digital asset holdings exceed 50% of total assets, aiming to reduce exposure to volatile cryptocurrencies in passive investment vehicles. This move targets major Bitcoin treasury holders like Strategy, potentially disrupting billions in investment flows. Strive Enterprises, a key player in the space, has formally opposed it through a letter to MSCI’s leadership. How Does the MSCI Bitcoin Exclusion Affect Bitcoin Treasury Firms? The proposal could deliver a substantial setback to Bitcoin treasury firms by limiting their inclusion in widely tracked MSCI indexes, which guide trillions in passive investments globally. According to JPMorgan analysts, Strategy alone might see a $2.8 billion drop in assets under management if excluded from the MSCI World Index, as reported in their recent market analysis. This exclusion would hinder these firms’ ability to attract institutional capital, forcing them to compete at a disadvantage against traditional finance entities. Strive CEO Matt Cole, in his letter to…
Share
BitcoinEthereumNews2025/12/06 11:33
Snowflake and Anthropic Forge $200M AI Partnership for Global Enterprises

Snowflake and Anthropic Forge $200M AI Partnership for Global Enterprises

The post Snowflake and Anthropic Forge $200M AI Partnership for Global Enterprises appeared on BitcoinEthereumNews.com. Peter Zhang Dec 04, 2025 16:52 Snowflake and Anthropic unveil a $200 million partnership to integrate AI capabilities into enterprise data environments, enhancing AI-driven insights with Claude models across leading cloud platforms. In a strategic move to enhance AI capabilities for global enterprises, Snowflake and Anthropic have announced a significant partnership valued at $200 million. This multi-year agreement aims to integrate Anthropic’s Claude models into Snowflake’s platform, offering advanced AI-driven insights to over 12,600 global customers through leading cloud services such as Amazon Bedrock, Google Cloud Vertex AI, and Microsoft Azure, according to Anthropic. Expanding AI Capabilities This collaboration marks a pivotal step in deploying AI agents across the world’s largest enterprises. By leveraging Claude’s advanced reasoning capabilities, Snowflake aims to enhance its internal operations and customer offerings. The partnership facilitates a joint go-to-market initiative, enabling enterprises to extract insights from both structured and unstructured data while adhering to stringent security standards. Internally, Snowflake has already been utilizing Claude models to boost developer productivity and innovation. The Claude-powered GTM AI Assistant, built on Snowflake Intelligence, empowers sales teams to centralize data and query it using natural language, thereby streamlining deal cycles. Innovative AI Solutions for Enterprises Thousands of Snowflake customers are processing trillions of Claude tokens monthly via Snowflake Cortex AI. The partnership’s next phase will focus on deploying AI agents capable of complex, multi-step analysis. These agents, powered by Claude’s reasoning and Snowflake’s governed data environment, allow business users to ask questions in plain English and receive accurate answers, achieving over 90% accuracy on complex text-to-SQL tasks based on internal benchmarks. This collaboration is especially beneficial for regulated industries like financial services, healthcare, and life sciences, enabling them to transition from pilot projects to full-scale production confidently. Industry Impact and Customer…
Share
BitcoinEthereumNews2025/12/06 11:17