Windows AI Foundry – Microsoft’s unified local AI development platform

What is Windows AI Foundry?

Windows AI Foundry is Microsoft’s newly launched AI development platform, designed to provide end-to-end support for AI developers—from model selection, optimization, and fine-tuning to deployment. Built upon the Windows Copilot Runtime and enhanced with new capabilities, the platform unifies tools and services across client and cloud environments.

Windows AI Foundry integrates Windows ML to enable efficient model deployment across various hardware (CPU, GPU, and NPU), and leverages Foundry Local to provide an optimized open-source model library for easy integration. It also offers a suite of ready-to-use AI APIs for tasks like text intelligence and image description. With LoRA (Low-Rank Adaptation) support, developers can fine-tune models like Phi Silica using custom data, improving development flexibility and efficiency.

Key Features of Windows AI Foundry

• Unified AI Development Platform
  Provides full lifecycle support—from model selection and optimization to fine-tuning and deployment—across both client and cloud environments.

• Windows ML Integration
  A built-in AI inference runtime that enables efficient model deployment on various hardware platforms including CPU, GPU, and NPU.

• Foundry Local
  Includes a rich library of open-source models. Automatically detects device hardware and recommends compatible models. Supports rapid model browsing, testing, and deployment.

• Ready-to-Use AI APIs
  Offers a variety of prebuilt AI APIs for text intelligence (e.g., summarization, rewriting) and image tasks (e.g., captioning, OCR), which can be executed locally without building or deploying models from scratch.

• LoRA Support
  Enables fine-tuning of the built-in Phi Silica model using LoRA, allowing developers to update only a small number of parameters with custom data to enhance model performance.

• Model Context Protocol (MCP) Integration
  Natively supports MCP, providing a standardized framework for connecting AI agents with local Windows applications. Developers can expose app functionality to agents, enhancing agent capabilities.

• App Actions on Windows
  Allows developers to create app-specific actions, increasing app discoverability and boosting user engagement.

Official Website

• Website: https://developer.microsoft.com/en-us/windows/ai/

Application Scenarios for Windows AI Foundry

• AI Beginners
  Easily browse and test pre-trained models and integrate them directly into local apps without the need for model training from scratch.

• Advanced Developers
  Optimize and deploy custom models across different hardware platforms, and fine-tune models using proprietary data to meet specialized needs.

• Application Developers
  Add intelligent features (e.g., text processing, image recognition) to applications, improving user experience and app visibility.

• Enterprise IT
  Leverage Windows’ built-in security features to protect data privacy during AI development and deployment.

• AI Agent Developers
  Build AI agents that interact with local Windows applications, expanding the capabilities and use cases of AI agents.