Microsoft Agent Framework Samples

A comprehensive hands-on guide to building intelligent agents using the Microsoft Agent Framework. This repository contains practical examples, tutorials, and code samples that demonstrate how to create powerful AI agents using both Python and .NET implementations.

🚀 What You'll Learn

This repository provides step-by-step tutorials and real-world examples covering:

• Agent Foundations: Core concepts and architecture of the Microsoft Agent Framework
• Creating Your First Agent: Build a simple travel planning agent from scratch
• Framework Exploration: Deep dive into different providers and configurations
• Tools Integration: Implement vision, code interpretation, and custom tools
• Provider Patterns: Work with MCP (Model Context Protocol) and Agent-to-Agent communication
• RAG Implementation: Build knowledge-enhanced agents with file search capabilities
• Multi-Agent Systems: Orchestrate multiple agents working together
• Workflow Management: Create complex agent workflows and pipelines

📁 Repository Structure

Directory	Description	.NET Code Samples	Python Code Samples
00.ForBeginners	Beginner-friendly Microsoft Agent Framework examples extending AI Agents for Beginners	Travel Agent Basic Agent Design Patterns Tool Use RAG Search Planning Multi-Agent	Travel Agent Basic Agent Design Patterns Tool Use RAG Search Planning Multi-Agent
01.AgentFoundation	Core concepts and architecture of Microsoft Agent Framework	Documentation Only	Documentation Only
02.CreateYourFirstAgent	Build your first travel planning agent from scratch	Travel Agent with GitHub Models	Travel Agent with GitHub Models
03.ExploreAgentFramework	Deep dive into different providers and configurations	Azure OpenAI GitHub Models MS Foundry Foundry Local	Azure OpenAI GitHub Models MS Foundry Foundry Local
04.Tools	Vision, code interpretation, and custom tool integration	Vision Code Interpreter Bing Grounding File Search	Vision Code Interpreter Bing Grounding File Search
05.Providers	MCP (Model Context Protocol) and Agent-to-Agent communication	MCP with Microsoft Learn	MCP with Microsoft Learn
06.RAGs	Knowledge-enhanced agents with file search capabilities	File Search RAG	File Search RAG
07.Workflow	Complex agent workflows and orchestration patterns	Basic Workflow Sequential Concurrent Conditional(MS Foundry)	Basic Workflow Sequential Concurrent Conditional(MS Foundry)
08.EvaluationAndTracing	Agent evaluation, debugging, and observability tools	GitHub Models Workflow DevUI	Single MS Foundry Agent DevUI Multi-Agent GitHub Models DevUI Multi-Agent MS Foundry DevUI Multi-Agent Foundry Local DevUI
09.Cases	Real-world case studies combining Foundry workflows with production-ready multi-agent applications	Microsoft Foundry with AITK & MAF GHModel Multi-Agent (.NET)	Microsoft Foundry with AITK & MAF GHModel Multi-Agent (Python) Agentic Marketing Content Generation

🛠 Prerequisites

Note: This is installation guideline

⚠️ Important Notice: Microsoft Agent Framework is currently in the development/preview stage. Since the framework APIs and features may change frequently, we strongly recommend building from source rather than using NuGet packages to ensure you have the latest updates and bug fixes.

📌 Additional Notes:

1. The examples in this repository are primarily based on GitHub Models and Microsoft Foundry. You can access GitHub Models directly at https://gh.io/models
2. Microsoft Foundry Agent Service uses the V2 version. Please access it at https://ai.azure.com and select V2 when creating or managing your Microsoft Foundry projects.

Python Environment

• Python 3.10 or higher
• Install dependencies:

pip install -r ./Installation/requirement.txt -U

Build from Source (Recommended):

git clone https://github.com/microsoft/agent-framework.git
cd agent-framework/python
pip install -e .

.NET Environment

• .NET 9.0 or higher
• Visual Studio 2022 or VS Code with C# extension

Build from Source (Recommended):

git clone https://github.com/microsoft/agent-framework.git
cd agent-framework/dotnet && dotnet build agent-framework-dotnet.slnx

After building, reference the local project in your notebooks or applications instead of NuGet packages. This ensures compatibility with the latest framework changes.

💻 Platform-Specific Setup

Windows ARM64 Configuration

If you're running on Windows ARM64, you may need to configure OpenSSL for certain dependencies:

git config --global core.longpaths true
winget install ShiningLight.OpenSSL.Dev
$env:OPENSSL_DIR="C:\Program Files\OpenSSL-Win64-ARM"
$env:OPENSSL_LIB_DIR="C:\Program Files\OpenSSL-Win64-ARM\lib\VC\arm64\MT"
$env:OPENSSL_STATIC="1"

Cross-Platform Compatibility

• Linux: Standard pip installation
• macOS: Homebrew for system dependencies
• Windows x64: Standard Windows installation

Required Services

• Azure OpenAI Service and Microsoft Foundry
• GitHub Models (for some examples)
• Azure CLI (authenticated)
• Azure Developer CLI (authenticated)

🚀 Quick Start

Environment Setup

Create a .env file in the root directory with your configurations:

GITHUB_TOKEN="Your GitHub Models Token"
GITHUB_ENDPOINT="Your GitHub Models Endpoint"
GITHUB_MODEL_ID="Your GitHub Model ID"

AZURE_OPENAI_ENDPOINT="Your Azure OpenAI Endpoint"
AZURE_OPENAI_CHAT_DEPLOYMENT_NAME ="Your Azure OpenAI Model Deployment Name"


FOUNDRYLOCAL_ENDPOINT="Your Foundry Local Endpoint http://localhost:5272/v1"
FOUNDRYLOCAL_MODEL_DEPLOYMENT_NAME="Your Foundry Local Model Deployment Name"


AZURE_AI_PROJECT_ENDPOINT ="Your Azure AI Foundry Project Endpoint"
AZURE_AI_MODEL_DEPLOYMENT_NAME ="Your Azure AI Foundry Project Deployment Name"

BING_CONNECTION_ID="Your Bing Connection ID"
BING_CONNECTION_NAME="Your Bing Connection Name"

OTEL_EXPORTER_OTLP_ENDPOINT="Your OpenTelemetry Collector Endpoint e.g. http://localhost:4317"

📚 Tutorial Progression

Getting Started Level

0. 00.ForBeginners - Comprehensive beginner tutorials with Microsoft Agent Framework examples

Foundation Level

1. 01.AgentFoundation - Understand the core concepts and architecture
2. 02.CreateYourFirstAgent - Build your first travel planning agent

Intermediate Level

3. 03.ExploreAgentFramework - Explore different providers (Azure OpenAI, GitHub Models, AI Foundry)
4. 04.Tools - Add vision, code interpretation, and custom tool capabilities
5. 06.RAGs - Implement knowledge-enhanced agents with file search

Advanced Level

6. 05.Providers - Master MCP (Model Context Protocol) and Agent-to-Agent communication
7. 07.Workflow - Create complex agent workflows and orchestration patterns
8. 08.EvaluationAndTracing - Learn evaluation, debugging, and observability tools for agents

🔧 Key Features Demonstrated

• Multiple Provider Support: Azure OpenAI, GitHub Models, Microsoft Foundry
• Tool Integration: Vision analysis, code interpretation, custom functions
• RAG Capabilities: File search and knowledge base integration
• Multi-Agent Orchestration: Sequential and collaborative agent patterns
• MCP Integration: Model Context Protocol for enhanced capabilities
• Streaming Responses: Real-time agent interactions
• Persistent Agents: Stateful agent conversations
• Evaluation & Debugging: DevUI for visual debugging and observability tools for tracing

🤝 Contributing

We welcome contributions! Please feel free to submit issues, feature requests, or pull requests.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🔗 Resources

• Microsoft Agent Framework Documentation
• Azure AI Services
• Microsoft Foundry
• GitHub Models

🆘 Support

If you encounter any issues or have questions:

1. Check the individual README files in each chapter directory
2. Review the code samples for implementation details
3. Open an issue in this repository
4. Consult the official Microsoft Agent Framework documentation

---

Start your journey with Microsoft Agent Framework today! 🚀