Home > Expertise > Microsoft Foundry

Microsoft Foundry

Connect with us on GitHub
Github
EXPERTISE
SHARE
EXPERTISE
SHARE
Unified Enterprise AI Platform for Building and Scaling Intelligent Agents

Microsoft Foundry is the comprehensive AI development platform empowering over 80,000 enterprises including 80% of Fortune 500 companies to build, deploy, and govern intelligent agents at scale. Launched at Microsoft Ignite 2025, it unifies access to 11,000+ AI models from Azure OpenAI, Anthropic Claude, Meta, Mistral, and more through serverless or managed compute. With Foundry Agent Service now generally available, developers can create multi-agent orchestration workflows using Agent2Agent protocol and Model Context Protocol for seamless integration with AutoGen, LangChain, and Semantic Kernel frameworks.

  • 11,000+ Frontier AI Models
    Access OpenAI (GPT-4o, GPT-5.2, o1), Anthropic Claude (Sonnet 4.5, Opus 4.1, Haiku 4.5), xAI, Meta, Mistral AI, Black Forest Labs, Cohere, and Microsoft Research models through unified platform.
  • Only Cloud with OpenAI & Anthropic
    Azure is the only cloud platform offering both OpenAI and Anthropic models with enterprise security, compliance, and regional availability for diverse AI workload requirements.
  • Foundry Agent Service (GA)
    Generally available agent service for building, deploying, and governing intelligent agents with multi-agent orchestration, Agent2Agent protocol, and Model Context Protocol (MCP) support.
  • Multi-Agent Orchestration
    Create complex agentic workflows using Agent2Agent protocol with seamless integration to AutoGen, LangChain, Semantic Kernel, and Copilot Studio frameworks for enterprise agent development.
  • Model Router with Auto-Selection
    Automatically selects optimal Azure OpenAI models for different prompts based on complexity, quality requirements, and cost constraints – boosting quality while reducing expenses.
  • Foundry IQ Powered by AI Search
    Streamlines how agents access and reason over public and private sources including SharePoint, Microsoft Fabric, web content, and enterprise data with grounded retrieval.
  • 1,000+ Tools & Connectors
    Foundry Tools catalog provides pre-built connectors to enterprise systems, APIs, databases, and services enabling rapid agent integration with existing business applications.
  • Enterprise Governance & Observability
    Organization-wide observability, control planes, compliance monitoring, and security governance built-in for responsible AI deployment at scale across departments.
  • Serverless or Managed Compute
    Deploy models with serverless pay-per-token pricing or dedicated managed compute for predictable performance and costs based on workload requirements and SLA needs.
  • 80,000+ Enterprise Customers
    Trusted by over 80,000 enterprises including 80% of Fortune 500 companies with proven scale, reliability, and enterprise-grade support for mission-critical AI applications.
Demos, docs and playgrounds

  • Access Microsoft Foundry portal to explore 11,000+ models, build agents with Agent Service, and deploy AI applications with enterprise governance.

    Microsoft Foundry Portal

  • Browse comprehensive catalog of AI models from OpenAI, Anthropic, Meta, Mistral, and more with model cards, benchmarks, and pricing information.

    Foundry Model Catalog

  • Interactive playground to test GPT-4o, Claude Sonnet 4.5, and other models with adjustable parameters, prompt engineering, and real-time responses.

    Azure AI Foundry Playground

  • Official GitHub repository with agent service examples, multi-agent orchestration patterns, and integration templates for enterprise development.

    Foundry Agent Service Samples
Helpful resources
  • Play
    Explore the new Microsoft Foundry portal
    Join April as she provides a tour of the new Microsoft Foundry portal experience. Learn how to access the new experience and explore several of the key areas to jump-start your agent building!