This sample showcases how to generate and orchestrate agents (aided by spec-kit) using Microsoft Agent Framework. It combines the best of:

• Semantic Kernel: Enterprise-ready AI orchestration
• AutoGen: Multi-agent conversation patterns

The sample is built around an engaging event planning scenario. It demonstrates concurrent workflow execution and real-time workflow visualization.

Plan a comprehensive event using multiple specialized agents working together.

• Event Coordinator: Orchestrates the overall planning process
• Venue Specialist: Researches and recommends venues
• Budget Analyst: Manages costs and financial constraints
• Catering Coordinator: Handles food and beverage planning
• Logistics Manager: Coordinates schedules and resources

1. Spec-to-Agent Generation: Create specialized agents by using spec-driven development with spec-kit
2. Event Planning Orchestration: Comprehensive event planning with multiple collaborating agents
3. Concurrent Workflows: Agent orchestration with parallel execution paths and fan-in/fan-out patterns
4. Interactive Frontend: Visual representation of agent interactions and workflows with DevUI
5. Azure Integration: Full AZD template deployment to Azure Container Apps

• Microsoft Agent Framework: Comprehensive multi-language framework for building, orchestrating, and deploying AI agents with support for both .NET and Python
• DevUI: Interactive UI showing real-time agent orchestration inspired by DevUI in Agent Framework
• Concurrent Workflows: Demonstrates the workflows functionality for orchestrating complex agentic systems
• Azure Deployment: Complete AZD template for one-click deployment

• Python 3.13+
• uv (Python package manager)
• Azure Developer CLI (azd)
• Azure subscription with appropriate permissions
• Node.js 18+ and npm

The easiest way to deploy this sample is using azd:

# Clone the repository
git clone https://github.com/microsoft/spec-to-agents.git
cd spec-to-agents

# Authenticate with Azure
azd auth login

# Initialize and deploy everything
azd up

When prompted:

• Environment name: Choose a name (e.g., dev, prod)
• Azure location: Choose from eastus2, westus, westus2, westus3, eastus, uksouth, swedencentral, australiaeast, or japaneast
• Azure subscription: Select your subscription

What azd up does:

• Provisions Azure resources (AI Foundry, Cosmos DB, Storage, App Services, etc.)
• Automatically generates a .env file with infrastructure outputs (via postprovision hook)
• Builds and deploys the backend API
• Builds and deploys the frontend application
• Configures all connections and settings

Alternative commands:

# Just provision infrastructure
azd provision

# Just deploy apps (after provisioning)
azd deploy

# View deployed endpoints
azd show

# Set environment variables
azd env set AZURE_LOCATION eastus2

Follow the development setup instructions here.

Note: Local development requires an Azure AI Foundry project and Azure OpenAI deployment. Update backend/.env with your Azure credentials.

• Backend: Python 3.13, FastAPI for High Performance ASGI Server
• Frontend: Vite frontend based on Microsoft Agent Framework DevUI
• Infrastructure: Azure Container Apps, AZD template
• AI: Azure OpenAI, Microsoft Agent Framework

1. Spec-First: Define agent specifications and capabilities using spec-kit.
2. Agent Generation: Create agents from specifications using the framework
3. Orchestration: Implement concurrent workflow and coordination patterns
4. Visualization: Build interactive frontend to show agent interactions
5. Deployment: Use AZD for one-click Azure deployment

cd tests
uv run pytest

The infrastructure is defined in the infra/ directory using Azure Bicep templates. It includes:

• Azure AI Foundry: AI Account and Project for agent orchestration
• Azure Cosmos DB: NoSQL database for agent state and memory
• Azure Storage: Blob storage for artifacts and data
• Azure AI Search: Vector search capabilities
• Azure App Services: Hosting for backend API and frontend
• Azure Application Insights: Monitoring and telemetry

All resources are provisioned automatically when you run azd provision or azd up.

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA).

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