Microsoft Foundry Bootstrap

A comprehensive toolkit for batch creation, testing, and management of AI agents at scale using Microsoft Foundry Control Plane features. Supports industry-specific templates, parallel agent operations, and real-time metrics collection through both a modern Web UI and an intuitive Terminal UI.

Screenshots

Web UI

Batch Agent Creation	Batch Workflow Creation

Agent Calling Simulation

Terminal UI

Agents	Workflows

Evaluations	Simulate

Features

• Dual Interface: Choose between modern Web UI (React) or Terminal UI (Textual) based on your preference
• Batch Agent Operations: Create and manage 100+ AI agents simultaneously with multi-threaded processing
• Industry Templates: Pre-configured profiles for Retail, Financial Services, Healthcare, Manufacturing, Logistics & Transportation, Energy & Utilities, and Telecommunications with specialized agent types
• Model Management: Discover existing models and deploy new ones through Microsoft Foundry Control Plane
• Workflow Management: Create and manage multi-agent workflows with configurable patterns
• Real-time Simulation: Run continuous daemon simulations with live metrics and progress tracking
• Code Generation: Automatically generate production-ready simulation scripts from templates
• Metrics & Visualization: Real-time performance tracking with interactive charts and dashboards
• Agent Registry: Centralized tracking of created agents with CSV-based persistence
• Sample Evaluations: Run reusable evaluation templates against selected agents using Foundry evaluation APIs

User Interfaces

Web UI (Recommended)

A modern, responsive web interface built with React, TypeScript, and Tailwind CSS.

Features:

• Visual dashboard with real-time status updates
• Interactive charts for simulation metrics
• Drag-and-drop workflows (coming soon)
• Progress tracking with live updates
• Dark mode support

Starting the Web UI:

# Start both API and frontend together
./start_all.sh

# Or start separately:
python start_api.py          # API server on http://localhost:8000
./start_frontend.sh          # Frontend on http://localhost:5173

Navigation:

• Dashboard - Overview of models, agents, workflows, and daemon status
• Templates - Browse and explore industry templates
• Agents - Create and manage AI agents with progress tracking
• Workflows - Create multi-agent workflows
• Simulation - Run one-time simulations or start continuous daemon
• Evaluations - Run evaluation templates against agents
• Settings - View models, connection status, and theme settings

Terminal UI (TUI)

A powerful command-line interface built with Textual for terminal-based management.

Features:

• Full keyboard navigation
• ASCII-based visualizations
• Works over SSH
• Low resource usage
• Scriptable operations

Starting the Terminal UI:

# Launch Terminal UI (default)
python main.py

# Or explicitly
python main.py tui

Terminal UI Screens:

• Main dashboard with agent/workflow counts
• Agent creation with template selection
• Model discovery and management
• Workflow creation
• Simulation runner
• Evaluation templates

CLI Commands (No UI)

For automation and scripting:

# List available templates
python main.py list

# Create agents from template
python main.py create retail -n 2 --orgs 3 -y

# Generate simulation code
python main.py generate retail -o output/retail_code

Quick Start

Prerequisites

• Python 3.10+
• Node.js 18+ (for Web UI)
• Azure CLI
• Microsoft Foundry project with endpoint

Setup

# 1. Create and activate virtual environment
python -m venv .venv
source .venv/bin/activate  # Linux/Mac
# .venv\Scripts\activate   # Windows

# 2. Install Python dependencies
pip install -r requirements.txt

# 3. Install frontend dependencies (for Web UI)
cd frontend
npm install
cd ..

# 4. Authenticate with Azure CLI (required for all operations)
az login

# 5. Configure environment
cp .env.example .env
# Edit .env and set your PROJECT_ENDPOINT

Launch Options

Option 1: Web UI (Recommended)

./start_all.sh
# Access at http://localhost:5173
# API docs at http://localhost:8000/docs

Option 2: Terminal UI

python main.py

Option 3: CLI Commands

python main.py list                    # List templates
python main.py create retail -n 2 -y   # Create agents

Project Structure

.
├── main.py                 # CLI entry point
├── start_api.py           # FastAPI server startup
├── start_frontend.sh      # React frontend startup
├── start_all.sh           # Start both services
├── requirements.txt        # Python dependencies
├── .env.example           # Environment template
│
├── src/                  # Core library
│   ├── api/             # FastAPI backend
│   │   ├── main.py            # FastAPI app
│   │   ├── routers/           # API route modules
│   │   ├── schemas/           # Request/response models
│   │   └── websocket.py       # WebSocket manager
│   ├── core/            # Business logic
│   │   ├── azure_client.py
│   │   ├── agent_manager.py
│   │   ├── workflow_manager.py
│   │   ├── model_manager.py
│   │   ├── evaluation_engine.py
│   │   ├── simulation_engine.py
│   │   └── daemon_service.py
│   ├── models/          # Pydantic models
│   ├── templates/       # Template utilities
│   └── codegen/         # Code generation
│
├── frontend/            # React Web UI
│   ├── src/
│   │   ├── components/        # Reusable UI components
│   │   ├── pages/            # Page components
│   │   ├── lib/              # API client and utilities
│   │   └── types/            # TypeScript types
│   ├── package.json
│   └── ...
│
├── templates/            # Industry templates
│   ├── industries/      # YAML profiles
│   │   ├── retail.yaml
│   │   ├── financial_services.yaml
│   │   ├── healthcare.yaml
│   │   └── manufacturing.yaml
│   └── code/            # Jinja2 templates
│
├── evaluation-templates/ # Sample evaluation templates (YAML)
│
└── ui/                   # Terminal UI
    ├── terminal/        # Textual TUI
    └── shared/          # Shared state

Industry Templates

Industry	Agent Types
retail	CustomerSupport, CatalogEnrichment, PricingOptimization, SupplyChain, MarketingCopy
financial_services	FraudDetection, RiskAssessment, ComplianceReview, CustomerService, PortfolioAnalysis
healthcare	PatientIntake, ClinicalDecisionSupport, BillingAssistant, AppointmentScheduler
manufacturing	QualityControl, PredictiveMaintenance, SupplyChainOptimizer, SafetyCompliance
logistics_transportation	DispatchCoordinator, RouteOptimizer, FleetManager, ShipmentTracking
energy_utilities	GridDispatcher, OutageCoordinator, LoadForecaster, MarketOperations
telecommunications	NetworkMonitoring, IncidentManager, ProvisioningSpecialist, FraudPrevention

API Endpoints

The FastAPI backend provides REST and WebSocket endpoints:

REST API

• GET /api/status - System status
• GET /api/models - List available models
• GET /api/templates - List industry templates
• GET /api/agents - List agents
• POST /api/agents - Create agents
• DELETE /api/agents - Delete all agents
• GET /api/workflows - List workflows
• POST /api/workflows - Create workflows
• DELETE /api/workflows - Delete all workflows
• POST /api/simulations/start - Start simulation
• POST /api/daemon/start - Start daemon
• POST /api/daemon/stop - Stop daemon
• POST /api/evaluations/run - Run evaluations

WebSocket Endpoints

• ws://localhost:8000/ws/simulation - Real-time simulation updates
• ws://localhost:8000/ws/daemon - Real-time daemon metrics

API Documentation: http://localhost:8000/docs

Environment Setup

Create .env from the example:

PROJECT_ENDPOINT=https://your-project.services.ai.azure.com/api/projects/your-project

Authentication uses Azure CLI credentials via DefaultAzureCredential, so run az login before using the CLI or TUI.

Usage Tips

Web UI Best Practices

• Dashboard: Monitor real-time status of models, agents, workflows, and daemon
• Progress Tracking: All create/delete operations show live progress bars
• Chart Styles: Switch between dot, line, and curve styles in simulation charts
• Dark Mode: Toggle theme in Settings page
• Deletion: Use "Delete All" with caution - operations cannot be undone

Terminal UI Navigation

• Use arrow keys or j/k to navigate
• Tab to switch between panels
• Enter to select
• q to quit most screens
• Ctrl+C to exit application

Performance Considerations

• Batch Operations: Creating 100+ agents uses multi-threading for optimal performance
• Daemon Mode: Runs continuous simulations with configurable intervals (60s default)
• Progress Callbacks: Real-time updates don't significantly impact operation speed
• WebSocket: Efficient real-time updates without polling overhead

Troubleshooting

Authentication Issues:

az login
az account show  # Verify correct subscription

Port Conflicts:

# Change ports in start_api.py or start_frontend.sh
uvicorn src.api.main:app --host 0.0.0.0 --port 8001
npm run dev -- --port 5174

Frontend Build Issues:

cd frontend
rm -rf node_modules package-lock.json
npm install
npm run dev

Missing Dependencies:

pip install -r requirements.txt --upgrade
cd frontend && npm install

Development

Running Tests

# API unit tests
pytest tests/unit/api/ -v

# All tests
pytest tests/ -v

Project Commands

# Format code
black src/
isort src/

# Type checking
mypy src/

# Linting
ruff check src/

License

[Your License Here]