MicrosoftDocs · qiutongMS · Apr 24, 2026 · Apr 24, 2026
diff --git a/docs/apis/get-started.md b/docs/apis/get-started.md
@@ -39,7 +39,7 @@ Ensure that your PC supports Windows AI APIs and that all dependencies are insta
 - Confirm that your device is a Copilot+ PC (we recommend the devices listed in the [Copilot+ PCs developer guide](../npu-devices/index.md).
 - Install [Windows 11 Insider Preview build 26120.3073 (Dev and Beta Channels)](https://blogs.windows.com/windows-insider/2025/01/31/announcing-windows-11-insider-preview-build-26120-3073-dev-and-beta-channels/), or later (to check your OS version, run `winver` from Windows Search).
 - Enable Developer Mode in **Settings** > **System** > **For developers** > **Developer Mode**.
-- Install [Visual Studio](https://visualstudio.microsoft.com/downloads/) with the specific workloads and components for developing with WinUI and the Windows App SDK. For more details, see [Required workloads and components](/windows/apps/get-started/start-here#22-required-workloads-and-components).
+- Install [Visual Studio](https://visualstudio.microsoft.com/downloads/) with the specific workloads and components for developing with WinUI and the Windows App SDK. For more details, see [Start here](/windows/apps/get-started/start-here).
 
 ---
 

diff --git a/docs/apis/image-generation.md b/docs/apis/image-generation.md
@@ -20,7 +20,7 @@ For **API details**, see [API ref for AI imaging features](/windows/windows-app-
 ## Prerequisites
 
 - **Windows version:** Windows 11, version 24H2 (build 26100) or later
-- **WinAppSDK version:** [Version 2.0 Experimental (2.0.0-Experimental3)](/windows/apps/windows-app-sdk/experimental-channel#version-20-experimental-200-experimental3)
+- **WinAppSDK version:** [Version 2.0 Experimental](/windows/apps/windows-app-sdk/experimental-channel)
 - **Hardware:** NPU-enabled PC recommended  
 
 ## What can I do with AI Image Generation?

diff --git a/docs/apis/index.md b/docs/apis/index.md
@@ -20,13 +20,13 @@ See the [Windows AI APIs with WinUI sample app](https://github.com/microsoft/Win
 > [!IMPORTANT]
 > The following is a list of Windows AI features and the Windows App SDK release in which they are currently supported. See [Overview of available APIs](#overview-of-available-apis) later in this topic for brief descriptions.
 >
-> [**Version 1.8.0 (1.8.250907003)**](/windows/apps/windows-app-sdk/stable-channel#version-18) - [Phi Silica (Limited Access Feature)](phi-silica.md), [Conversation Summarization (Text Intelligence)](phi-silica.md#text-intelligence-skills), [Object Erase](./image-object-erase.md)
+> [**Version 1.8.0 (1.8.250907003)**](/windows/apps/windows-app-sdk/stable-channel) - [Phi Silica (Limited Access Feature)](phi-silica.md), [Conversation Summarization (Text Intelligence)](phi-silica.md#text-intelligence-skills), [Object Erase](./image-object-erase.md)
 >
-> [**Version 1.8 Preview (1.8.0-preview)**](/windows/apps/windows-app-sdk/preview-channel#version-18-preview-18-preview) - [LoRA fine-tuning for Phi Silica](phi-silica-lora.md), [Text Rewriter Tone (Text Intelligence)](phi-silica.md#text-intelligence-skills)
+> [**Version 1.8 Preview (1.8.0-preview)**](/windows/apps/windows-app-sdk/preview-channel) - [LoRA fine-tuning for Phi Silica](phi-silica-lora.md), [Text Rewriter Tone (Text Intelligence)](phi-silica.md#text-intelligence-skills)
 >
 > [**Private preview**](https://aka.ms/WindowsAIFSemanticSearch) - Semantic Search
 >
-> [**Version 1.7.1 (1.7.250401001)**](/windows/apps/windows-app-sdk/stable-channel#version-171-17250401001) - All other APIs
+> [**Version 1.7.1 (1.7.250401001)**](/windows/apps/windows-app-sdk/downloads) - All other APIs
 
 ## Build your first AI-powered Windows app
 

diff --git a/docs/apis/phi-silica-tutorial.md b/docs/apis/phi-silica-tutorial.md
@@ -78,7 +78,7 @@ std::wcout << result.Text() << std::endl;
 ## Build and run the sample
 
 1. Clone the [WindowsAppSDK-Samples](https://github.com/microsoft/WindowsAppSDK-Samples) repo.
-1. Navigate to the [Samples/WindowsAIFoundry/cs-maui](https://github.com/microsoft/WindowsAppSDK-Samples/tree/main/Samples/WindowsAIFoundry/cs-maui) folder.
+1. Navigate to the [Samples/WindowsAIFoundry/cs-maui](https://github.com/microsoft/WindowsAppSDK-Samples/tree/release/experimental/Samples/WindowsAIFoundry/cs-maui) folder.
 1. Open MauiWindowsAISample.sln in Visual Studio 2022.
 1. Ensure the debug toolbar has "Windows Machine" set as the target device.
 1. Press F5 or select "Start Debugging" from the Debug menu to run the sample (the sample can also be run without debugging by selecting "Start Without Debugging" from the Debug menu or Ctrl+F5).

diff --git a/docs/apis/phi-silica-winui-tutorial.md b/docs/apis/phi-silica-winui-tutorial.md
@@ -38,7 +38,7 @@ A WinUI 3 app with:
 
 3. **Developer Mode** enabled — Windows Settings → System → For developers → Developer Mode.
 
-4. **Visual Studio 2022** with the **Windows application development** workload installed. See [Required workloads and components](/windows/apps/get-started/start-here#22-required-workloads-and-components).
+4. **Visual Studio 2022** with the **Windows application development** workload installed. See [Start here](/windows/apps/get-started/start-here).
 
 5. **Windows App SDK 2.0 preview NuGet package** — `Microsoft.WindowsAppSDK` version `2.0.0-preview1`. You'll install this in the steps below.
 

diff --git a/docs/apis/troubleshooting.md b/docs/apis/troubleshooting.md
@@ -39,7 +39,7 @@ Ensure that you can use Phi Silica features:
 
 Enable **Developer Mode** in **Settings** > **System** > **For developers** > **Developer Mode**.
 
-Install [Visual Studio](https://visualstudio.microsoft.com/downloads/) with the workloads and components required for developing with WinUI and the Windows App SDK. For more details, see [Required workloads and components](/windows/apps/get-started/start-here#22-required-workloads-and-components).
+Install [Visual Studio](https://visualstudio.microsoft.com/downloads/) with the workloads and components required for developing with WinUI and the Windows App SDK. For more details, see [Start here](/windows/apps/get-started/start-here).
 
 ## Debugging
 
@@ -67,15 +67,15 @@ If Phi Silica does not appear to be working on your system, there are a few thin
 
 ## Other guidance
 
-- Apps that use Windows AI APIs need to be granted package identity at runtime. For details, see [Advantages and disadvantages of packaging your app](/windows/apps/package-and-deploy/#advantages-and-disadvantages-of-packaging-your-app). If you are seeing an **UnauthorizedAccessException** error (or having other access issues), ensure that your app is packaged and the **systemAIModels** capability was added to your manifest file (see [Get started building an app with Windows AI APIs](./get-started.md)).
+- Apps that use Windows AI APIs need to be granted package identity at runtime. For details, see [Packaging overview](/windows/apps/package-and-deploy/packaging/). If you are seeing an **UnauthorizedAccessException** error (or having other access issues), ensure that your app is packaged and the **systemAIModels** capability was added to your manifest file (see [Get started building an app with Windows AI APIs](./get-started.md)).
 
 - Limited Access Feature (LAF) requirements:
   - Verify that you've requested and received a LAF token from Microsoft
   - The following status values indicate that LAF has failed.
     - [LimitedAccessFeatureStatus.Unavailable](/uwp/api/windows.applicationmodel.limitedaccessfeaturestatus)
     - [LimitedAccessFeatureStatus.Unknown](/uwp/api/windows.applicationmodel.limitedaccessfeaturestatus)
 
-- Self-contained apps cannot run from the **Downloads** folder (or from anywhere under the `C:\Users` folder). For more details, see [Advantages and disadvantages of packaging your app](/windows/apps/package-and-deploy/#advantages-and-disadvantages-of-packaging-your-app) and [Windows App SDK deployment overview](/windows/apps/package-and-deploy/deploy-overview).
+- Self-contained apps cannot run from the **Downloads** folder (or from anywhere under the `C:\Users` folder). For more details, see [Packaging overview](/windows/apps/package-and-deploy/packaging/) and [Windows App SDK deployment overview](/windows/apps/package-and-deploy/deploy-overview).
 
 - The Windows App SDK experimental channel includes APIs and features in early stages of development. All APIs in the experimental channel are subject to extensive revisions and breaking changes and may be removed from subsequent releases at any time. Experimental features are not supported for use in production environments and apps that use them cannot be published to the Microsoft Store.
 

diff --git a/docs/cards/image-ai-apis-platform-card.md b/docs/cards/image-ai-apis-platform-card.md
@@ -82,7 +82,7 @@ Image AI APIs are designed for a range of use cases across accessibility, creati
 
 ## Models and training data
 
-The Image AI APIs in Windows leverage machine learning models trained to perform perception tasks such as segmentation, object detection, image enhancement, and captioning. These proprietary models are developed and trained by Microsoft using a combination of licensed data, publicly available data, and human-annotated images, in accordance with Microsoft's Responsible AI standards. For more information on Microsoft's approach to Responsible AI development, please refer to this link: [Responsible Generative AI Development on Windows](/windows/ai/responsible-ai).
+The Image AI APIs in Windows leverage machine learning models trained to perform perception tasks such as segmentation, object detection, image enhancement, and captioning. These proprietary models are developed and trained by Microsoft using a combination of licensed data, publicly available data, and human-annotated images, in accordance with Microsoft's Responsible AI standards. For more information on Microsoft's approach to Responsible AI development, please refer to this link: [Developing Responsible Generative AI Development on Windows](/windows/ai/rai).
 
 ## Performance
 
@@ -166,12 +166,12 @@ Apply human oversight in final content use, especially when image descriptions a
 
 For additional guidance on the responsible use of Image AI APIs, we recommend reviewing the following documentation.
 
-- [Image Foreground Extractor API | Microsoft Learn](/windows/ai/image-foreground-extractor)
-- [Image Object Erase API | Microsoft Learn](/windows/ai/image-object-erase)
-- [Image Object Extractor API | Microsoft Learn](/windows/ai/image-object-extractor)
-- [Image Super-Resolution API | Microsoft Learn](/windows/ai/image-super-resolution)
-- [Image Description API | Microsoft Learn](/windows/ai/image-description)
-- [Responsible Generative AI Development on Windows | Microsoft Learn](/windows/ai/responsible-ai)
+- [Image Foreground Extractor API | Microsoft Learn](/windows/ai/apis/image-foreground-extractor)
+- [Image Object Erase API | Microsoft Learn](/windows/ai/apis/image-object-erase)
+- [Image Object Extractor API | Microsoft Learn](/windows/ai/apis/image-object-extractor)
+- [Image Super-Resolution API | Microsoft Learn](/windows/ai/apis/image-super-resolution)
+- [Image Description API | Microsoft Learn](/windows/ai/apis/image-description)
+- [Developing Responsible Generative AI Development on Windows | Microsoft Learn](/windows/ai/rai)
 
 **Learn more about responsible AI**
 

diff --git a/docs/cards/phi-silica-platform-card.md b/docs/cards/phi-silica-platform-card.md
@@ -95,7 +95,7 @@ For organizations or developers that load LoRA adapters, Phi Silica can adapt to
 
 ## Models and training data
 
-This model is developed and trained by Microsoft in accordance with Microsoft's Responsible AI standards. For more information on Microsoft's approach to Responsible AI development, please refer to this link: [Responsible Generative AI Development on Windows](/windows/ai/responsible-ai).
+This model is developed and trained by Microsoft in accordance with Microsoft's Responsible AI standards. For more information on Microsoft's approach to Responsible AI development, please refer to this link: [Responsible Generative AI Development on Windows](/windows/ai/rai).
 
 ## Performance
 
@@ -178,9 +178,9 @@ Responsible AI is a shared commitment between Microsoft and its customers. While
 
 For additional guidance on the responsible use of Phi Silica, we recommend reviewing the additional documentation below.
 
-- [Get started with Phi Silica in the Windows App SDK | Microsoft Learn](/windows/apps/windows-ai/phi-silica)
-- [Content safety moderation with the Windows AI APIs | Microsoft Learn](/windows/ai/content-safety)
-- [Responsible Generative AI Development on Windows | Microsoft Learn](/windows/ai/responsible-ai)
+- [Get started with Phi Silica in the Windows App SDK | Microsoft Learn](/windows/ai/apis/phi-silica)
+- [Content safety moderation with the Windows AI APIs | Microsoft Learn](/windows/ai/apis/content-moderation)
+- [Developing Responsible Generative AI Development on Windows | Microsoft Learn](/windows/ai/rai)
 
 **Learn more about responsible AI**
 

diff --git a/docs/cards/text-recognition-ocr-platform-card.md b/docs/cards/text-recognition-ocr-platform-card.md
@@ -72,7 +72,7 @@ Text Recognition (OCR) is designed for a broad range of productivity, accessibil
 
 ## Models and training data
 
-Text Recognition (OCR) in Windows is powered by Microsoft-developed machine learning models trained to recognize text in images. These models are trained using a combination of licensed data, data created by human trainers, and publicly available text-image data, in accordance with Microsoft's Responsible AI standards. For additional information on Microsoft's approach to responsible AI development, please refer to this link: [Responsible Generative AI Development on Windows](/windows/ai/responsible-ai).
+Text Recognition (OCR) in Windows is powered by Microsoft-developed machine learning models trained to recognize text in images. These models are trained using a combination of licensed data, data created by human trainers, and publicly available text-image data, in accordance with Microsoft's Responsible AI standards. For additional information on Microsoft's approach to responsible AI development, please refer to this link: [Responsible Generative AI Development on Windows](/windows/ai/rai).
 
 ## Performance
 
@@ -150,9 +150,9 @@ Responsible AI is a shared commitment between Microsoft and its customers. While
 
 For additional guidance, see the following resources:
 
-- [Text Recognition (OCR) API documentation | Microsoft Learn](/windows/ai/ocr)
-- [Text Recognition tutorial | Microsoft Learn](/windows/ai/ocr-tutorial)
-- [Responsible Generative AI Development on Windows | Microsoft Learn](/windows/ai/responsible-ai)
+- [Text Recognition (OCR) API documentation | Microsoft Learn](/windows/ai/apis/text-recognition)
+- [Text Recognition tutorial | Microsoft Learn](/windows/ai/apis/text-recognition-tutorial)
+- [Responsible Generative AI Development on Windows | Microsoft Learn](/windows/ai/rai)
 
 **Learn more about responsible AI**
 

diff --git a/docs/npu-devices/index.md b/docs/npu-devices/index.md
@@ -136,7 +136,7 @@ To perform these measurements, we recommend the following diagnostic and tracing
   - Learn more about [ONNX Runtime Execution Provider (EP) Profiling](https://onnxruntime.ai/docs/performance/tune-performance/profiling-tools.html#execution-provider-ep-profiling).
 
 > [!NOTE]
-> WPR UI (the user interface available to support the command-line based WPR included in Windows), WPA, and GPUView are all part of Windows Performance Toolkit (WPT), version May 2024+. To use the WPT, you will need to: [Download the Windows ADK Toolkit](/windows-hardware/get-started/adk-install#download-the-adk-101261001-may-2024).
+> WPR UI (the user interface available to support the command-line based WPR included in Windows), WPA, and GPUView are all part of Windows Performance Toolkit (WPT). To use the WPT, you will need to: [Download the Windows ADK Toolkit](/windows-hardware/get-started/adk-install#download-the-adk-101261002454-december-2024).
 
 For a quickstart on viewing ONNX Runtime events with the Windows Performance Analyzer (WPA), follow these steps: