Intelligent Azure architecture diagram generator using AI agents, MCP service validation, and comprehensive Azure component library.
🏗️ Architecture
This solution provides an intelligent system for generating professional Azure architecture documents and diagrams with automatic validation and correction:
AI-powered Azure Architecture Generator that automatically creates AI agents for design documents and diagram generation
✨ Features
🤖 AI-Powered Generation
Multi-Agent Architecture: Specialized agents for design, validation, and diagram generation
GPT-4o Integration: Advanced natural language processing for architecture understanding
Intelligent Component Selection: Automatic Azure service recommendations
🔧 MCP Validation Engine
Component Validation: Validates all Azure components against the official diagrams library
Automatic Correction: Fixes import paths and component names automatically
Smart Suggestions: Provides intelligent alternatives for invalid components
Comprehensive Coverage: Supports 247+ Azure services across all categories
📊 Professional Diagrams
High-Quality Rendering: Vector-based diagrams with professional styling
Multiple Formats: PNG, SVG, and PDF output support
Responsive Design: Works on desktop and mobile devices
Export Options: Download diagrams in various formats
🛡️ Enterprise-Ready
Managed Identity: Secure authentication without credentials
Cosmos DB Storage: Persistent architecture storage and versioning
Click the "Open in GitHub Codespaces" badge above or visit the repository
Create a new Codespace - The environment will automatically configure itself
Deploy to Azure with just two commands:
azd auth login
az login
azd up
No need to install anything locally - everything runs in the cloud! Perfect for developers who want to try the template quickly or work from any device.
Remember you should have logged in with Azure CLI or switched to your target subscription:
az login
az account set --subscription <your-subscription-id>
Option 2: Interactive Deployment (Recommended)
azd init --template passadis/ai-architect-webapp
az login
azd up
Option 3: Non-Interactive Deployment
az login
azd init --template passadis/ai-architect-webapp -e <environment-name> --up
azd up
The deployment will:
Create an Azure AI Foundry service with GPT-4o model
Deploy frontend, backend, and MCP service containers
Set up Cosmos DB, Storage Account, and Key Vault
Configure managed identity and RBAC permissions
Provide the application URL when complete
Azure AI Agent Service
This deployment uses the Basic Agent Setup which includes:
Azure AI Foundry service with Managed Identity authentication
AI project with GPT-4o model deployment
Microsoft-managed Key Vault
All necessary RBAC permissions
Alternative manual setup:
💡 Usage Examples - Prompts
Simple Web Application
Create a web application with a database and storage
Generated: App Service + SQL Database + Blob Storage with proper connections
Microservices Architecture
Design a microservices platform with API gateway, container apps, and shared database
Generated: API Management + Container Apps + Cosmos DB + Service Bus
Data Analytics Pipeline
Build a data lake solution with Azure Functions for processing and Power BI for visualization
Generated: Data Lake + Functions + Event Hub + Power BI with data flow connections
✨ What Makes This Project Special
🧠 Intelligent Architecture: Combines Azure AI automation with a clean, maintainable codebase
📘 Model Context Protocol (MCP): Acts as a single source of truth for validation and service integration
🔗 Seamless Azure Integration: Works effortlessly with Cosmos DB, Azure Storage, and other native services
⚙️ Scalable & Secure Backend: Designed for high performance and enterprise-grade security
🖥️ Intuitive Frontend: Responsive UI for a smooth and user-friendly experience
🧪 Clean Separation of Concerns: Modular design for easier testing, maintenance, and innovation
🚀 azd Template Support: Fully compatible with Azure Developer CLI for streamlined deployment
🏢 Enterprise Ready: Built for real-world cloud environments with rapid iteration in mind
🏛️ Architecture Components
Frontend (React + TypeScript)
Modern React application with TypeScript
Material-UI components for professional interface
Real-time diagram preview and editing
Responsive design for all devices
Backend API (FastAPI + Python)
FastAPI web framework for high-performance APIs
Azure AI Projects integration for GPT-4o access
Multi-agent orchestration system
Comprehensive error handling and logging
MCP Service (Python + MCP Protocol)
Model Context Protocol implementation
Azure components validation engine
Automatic import correction and suggestions
GraphViz integration for diagram rendering
Infrastructure (Terraform)
Azure Container Apps for scalable hosting
Cosmos DB for persistent storage
Azure AI Projects for AI capabilities
Managed Identity for secure authentication
🔧 Configuration
Environment Variables
The application uses the following environment variables (automatically configured during deployment):
Diagram Styling: Customize rendering in mcp-service/mcp_diagrams_server.py
Infrastructure: Modify Terraform files in infra/
🧪 Development
Cloud-Native Development (Recommended)
This project is designed as a cloud-native Azure template. For the best development experience:
Option 1: GitHub Codespaces
# One-click cloud development environment
# Click "Code" → "Create codespace on main" in GitHub
# All dependencies and Azure CLI pre-configured
azd auth login
az login
azd up
Option 2: Azure Deployment
# Deploy your development environment to Azure
azd auth login
az login or az account set --subscription <your-subscription-id>
azd init --template passadis/ai-architect-webapp
azd up
# Develop against live Azure services
Why No Local Full-Stack Setup?
This template integrates deeply with Azure services (AI Foundry, Cosmos DB, Managed Identity) that don't have local equivalents. The architecture is optimized for Azure Container Apps with Dapr service mesh.
For learning and experimentation, we recommend deploying to Azure's free tier or using GitHub Codespaces.