Snippy

What is Snippy?

Snippy is an Azure Functions-based reference application that demonstrates how to build MCP (Model Context Protocol) tools that integrate with AI assistants like GitHub Copilot. It showcases a serverless AI architecture using Azure Functions, Durable Functions, Azure OpenAI, and the Agent Framework for code snippet management and multi-agent orchestration.

How to use Snippy?

Deploy the entire stack to Azure with a single azd up command after satisfying prerequisites (Azure subscription, azd CLI, and a development environment such as GitHub Codespaces or VS Code Dev Containers). For local development, use Docker-based emulators and Azure Functions Core Tools v4. Once deployed, the Function App URL and MCP endpoint are displayed, ready for integration with AI assistants.

Key features of Snippy

• Expose Azure Functions as discoverable MCP tools.
• Stateful AI agents with automatic conversation history.
• Multi-agent orchestration via Durable Task Scheduler.
• Semantic code retrieval with Cosmos DB vector search.
• Real-time monitoring with DTS dashboard.
• One-click infrastructure deployment using azd up.

Use cases of Snippy

• Save code snippets with vector embeddings for semantic search.
• Retrieve snippets by name for quick reference.
• Generate language-specific code style guides from saved snippets.
• Create comprehensive wiki documentation via multi-agent workflow.
• Use as a reference architecture for building production‑ready MCP tools.

FAQ from Snippy

What is the purpose of Snippy?

Snippy is intended for learning and demonstration—it shows how to build MCP tools with Azure Functions. It should not be deployed to production without a thorough security review.

What are the prerequisites to deploy Snippy?

You need an Azure subscription with permission to create resources, the azd CLI installed, and a development environment (GitHub Codespaces, VS Code Dev Containers with Docker Desktop, or a local setup with Python 3.11, Node 18+, and Azure Functions Core Tools v4).

Which Azure regions are recommended?

The README recommends eastus or swedencentral because Azure OpenAI model availability varies by region. Verify model support in those regions before deploying.

How does Snippy handle security?

Snippy uses User-Assigned Managed Identity for service-to-service authentication, with RBAC roles for Cosmos DB, Storage, Application Insights, and Azure AI Project. For production, the README advises restricting network access via Private Endpoints and VNet integration.

Which MCP tools are included?

The tool matrix lists five tools: save_snippet, get_snippet, code_style, deep_wiki, and generate_comprehensive_documentation (an orchestrated multi-agent workflow).