modelmeters.com provides a simple front-end to the Azure Retail Prices API, automatically downloading and enriching Azure AI pricing data with AI-generated summaries and insights.
Model Meters combines several Azure and GitHub services to create an automated pricing intelligence solution:
- Daily Data Collection: GitHub Actions automatically downloads the latest Azure pricing data using the Azure Retail Prices API
- Data Processing: The raw pricing data is split into segments based on the pricing
startDatevalue - AI Enhancement: Azure OpenAI (via Azure AI Foundry) generates intelligent summaries of the latest pricing changes, using Model Context Protocol (MCP) to include Microsoft Learn documentation
- Web Deployment: The processed data and summaries are automatically deployed to Azure Static Web Apps using GitHub Actions
- User Interface: A responsive web interface allows users to explore pricing data and AI-generated insights
- GitHub Actions Workflow: Runs daily at midnight UTC to check for pricing updates
- AI-Powered Summaries: Leverages Azure OpenAI to generate concise, factual summaries of pricing changes grouped by model provider
- Documentation Integration: Uses Microsoft Learn MCP Server to automatically include relevant documentation links
- Human-in-the-loop: As AI summaries are stored in markdown format, this allows for quick and easy human review (and potentially AI agent review)
This sample is designed to help Microsoft partners (and their customers) understand how they can:
- Use GitHub and GitHub Copilot to rapidly build a simple AI agent for data processing
- Combine GitHub Actions with Azure AI Foundry to automate repeatable tasks at scale
- Use LLMs to enrich frequently updated structured data, such as price lists, inventory, or sales data
- Incorporate the use of Model Context Protocol (MCP) within their solutions for enhanced AI capabilities
- Build cost-effective monitoring solutions for tracking Azure service pricing changes
- Create automated reporting systems that combine real-time data with AI-generated insights
- GitHub Copilot - AI-powered code completion and development assistance
- GitHub Actions - CI/CD automation and workflow orchestration
- Azure Static Web Apps - Static site hosting with integrated CI/CD
- Azure AI Foundry - AI model deployment and management platform
- Model Context Protocol (MCP) - Protocol for AI model context sharing
- HTML/CSS/JavaScript - Frontend web interface
- Python 3.1x - Backend data processing and AI integration
- OpenAI Python SDK - AI model interaction
- JSON/NDJSON - Data storage and interchange formats
- Azure Retail Prices API - Official Azure pricing data source
- Azure OpenAI Service - AI model hosting and inference
- Azure Static Web Apps - Web application hosting and deployment
- Microsoft Learn MCP Server - Documentation context provider for AI summaries
- Responses API - AI response generation, with support for MCP
- GitHub Actions - Workflow automation and CI/CD
- GitHub Models - Alternative AI model access (fallback option, especially for simpler, smaller demos)
The minimal solution takes advantage of free tiers where available, and has been designed to be deployed on a Visual Studio Subscription, or an Azure Free account. The Advanced demo could be deployed on Azure Bulk Credit, or on a commercial Azure subscription (CSP, PAYG, MCA-E, etc.).
| Component | Minimal demo | Advanced demo | Notes |
|---|---|---|---|
| Azure Static Web Apps | $0/month | $9/month (Standard) | Free tier includes 100GB bandwidth, 0.5GB storage |
| Azure AI Foundry - OpenAI | ~$5-20/month | ~$50-100/month | Price varies based on token usage |
| Azure Retail Prices API | $0/month | $0/month | Free public API with rate limits |
| GitHub Actions | $0/month | $4/month (Team plan) | 2000 minutes/month free, then $0.008/minute |
| GitHub Repository | $0/month | $4/month (Team plan) | Public and private repos free, additional features in Team plan |
| Domain/Custom DNS | Optional $15/year | Optional $15/year | Optional custom domain, via third-party domain registrar |
| Monitoring/Analytics | $0/month | $10-25/month | Optional Application Insights, etc. |
| Storage (backup/logs) | $0-2/month | $5-15/month | Azure Storage for additional data retention |
| Azure AI Search | Optional (Free tier) | Optional ~$75/month (Basic tier) | Free tier includes up to 3 indexes, 50MB storage; Basic tier allows larger workloads |
| Microsoft Fabric | N/A | Optional ~$262/month (F2 capacity) | Based on Fabric F2 capacity |
| Total Monthly Cost | $5-10/month | $100-500/month |
Pricing Disclaimer: Costs are approximate, correct at time of writing (August 2025), and may vary based on:
- Actual usage patterns and data volumes
- Regional pricing differences
- Azure service tier selections
- Token consumption for AI summaries
- Additional monitoring or storage requirements
Other costs may be incurred depending on specific implementation choices and usage patterns.
- Azure subscription with AI Foundry access
- GitHub repository
- Python 3.11+ for local development
-
Clone the repository:
git clone https://github.com/guygregory/modelmeters.com.git cd modelmeters.com -
Configure Azure OpenAI (for AI summaries):
# Set environment variables export AZURE_OPENAI_API_KEY="your-api-key" export AZURE_OPENAI_V1_API_ENDPOINT="https://your-resource.openai.azure.com/openai/v1/" export AZURE_OPENAI_API_MODEL="gpt-4"
-
Configure GitHub Token (for GitHub Models fallback):
export GITHUB_TOKEN="your-github-token"
-
Download pricing data:
python meter-download.py --cognitive-services-only --ndjson prices.ndjson
-
Process monthly data:
python split_into_monthly.py
-
Generate AI summaries:
python create-ai-summaries.py
-
Serve locally:
python -m http.server 8000 # Visit http://localhost:8000
The project automatically deploys to Azure Static Web Apps via GitHub Actions when changes are pushed to the main branch.
├── index.html # Main pricing data explorer
├── agent/ # AI summary interface
├── monthly/ # Monthly pricing data and summaries
│ ├── full/ # Complete monthly data files
│ ├── partial/ # Filtered monthly data files
│ └── aisummary/ # AI-generated summaries
├── .github/workflows/ # GitHub Actions automation
├── ai-summary.py # Azure OpenAI summary generation
├── ai-summary-github-models.py # GitHub Models fallback
├── meter-download.py # Azure pricing data downloader
├── split_into_monthly.py # Data processing utilities
└── prices.ndjson # Latest pricing data
- Real-time filtering and search capabilities
- Sortable columns with customizable views
- Export functionality for data analysis
- Responsive design for mobile and desktop
- Daily automated summaries of pricing changes
- Model provider groupings and comparisons
- Integration with Microsoft Learn documentation
- Contextual links to relevant resources
- Daily data refresh via GitHub Actions
- Intelligent change detection to minimize costs
- Automatic deployment on data changes
- Comprehensive error handling and retry logic
# Basic pricing data retrieval
import requests
response = requests.get(
"https://prices.azure.com/api/retail/prices",
params={"$filter": "serviceName eq 'Cognitive Services'"}
)
pricing_data = response.json()# Generate AI summary using Azure OpenAI
from openai import OpenAI
client = OpenAI(
api_key="your-api-key",
base_url="https://your-resource.openai.azure.com/openai/v1/",
default_query={"api-version": "preview"}
)
response = client.responses.create(
model="gpt-4",
instructions="Summarize Azure pricing changes...",
tools=[{
"type": "mcp",
"server_label": "MicrosoftLearn",
"server_url": "https://learn.microsoft.com/api/mcp"
}],
input=pricing_data
)- Fork the repository
- Create a feature branch
- Make your changes
- Test locally using the setup instructions
- Submit a pull request
- Demonstration Purposes: The site and code in this repository are for demonstration purposes only, not intended for production use
- No Warranties: No warranties, guarantees, or support is provided for this code
- AI-Generated Content: Summaries are AI-generated and could contain mistakes or inaccuracies
- Authoritative Source: Always refer to the official Azure pricing page and price lists directly for authoritative pricing information
- Data Accuracy: While we strive for accuracy, pricing data may be delayed or incomplete
- Usage Responsibility: Users are responsible for validating any pricing information before making business decisions
- Azure Pricing Documentation: https://docs.microsoft.com/azure/cost-management-billing/
- Azure AI Foundry: https://docs.microsoft.com/azure/ai-foundry/
- GitHub Actions: https://docs.github.com/en/actions
- Azure Static Web Apps: https://docs.microsoft.com/azure/static-web-apps/
- Microsoft Learn MCP Server: https://github.com/microsoftdocs/mcp
This project is provided under the MIT License. See LICENSE file for details.