How hyssh/quickstart-sk works
Project type and purpose
• This is a full-stack AI orchestration platform. It uses FastAPI as a backend “semantic kernel” orchestrator, Chainlit as an interactive chat UI frontend, and a set of MCP (Model Context Protocol) microservices for specialized data enrichment (weather, local time, Azure Data Explorer). It also includes sample MCP services for Azure Search and Azure SQL.
File structure
• Top level:
• .env.example, requirements.txt, .gitignore – configuration and dependencies.
• sample-server/ – example MCP implementations for Azure Search and SQL DB.
• src/ – main application code:
– launcher.py & run_chainlit.cmd – unified entrypoint for starting all components.
– backend/server.py – FastAPI backend.
– frontend/app.py, chainlit.yaml, chainlit.md – Chainlit chat UI.
– mcpservers/*.py – independent microservices (weather, localtime, azuredataexplorer).
README for additional insights
• The README spells out five runtime services:
1. Weather MCP server
2. LocalTime MCP server
3. Azure Data Explorer MCP server
4. FastAPI backend server
5. Chainlit frontend
• It describes environment variables for Azure AI Foundry, search, ADX, and the launcher patterns.
Detail
• Ports (e.g., backend: 8091, Chainlit: 8000) and environment variables.