ChatWithDb

Interact with your own MongoDB data through a conversational interface powered by Semantic Kernel. This portfolio-ready project pairs an ASP.NET Core backend with a React frontend to deliver read-only analytics, table views, and charts generated by an LLM.

Features

• Connect to any MongoDB instance via connection string (locally stored per session).
• Chat with an LLM that can call vetted functions for schema discovery and query execution.
• Built-in Semantic Kernel plugin for MongoDB collection introspection and safe aggregation.
• Render tabular results and data visualizations based on query outputs.
• Enforced read-only access with guardrails for safe tool execution.
• Modular architecture ready for future providers or databases.

Project Structure

.
├── backend/      # ASP.NET Core minimal API + Semantic Kernel plugins
├── frontend/     # React (Vite + TypeScript) UI
├── README.md     # Project overview and setup guide
└── .gitignore

Tech Stack

• Backend: .NET 8, ASP.NET Core Minimal APIs, Microsoft Semantic Kernel, MongoDB Driver
• Frontend: React, Vite, TypeScript, TailwindCSS (planned), recharts
• LLM Providers: Azure OpenAI (default) with Semantic Kernel abstraction for other providers

Getting Started

1. Clone the repository and install prerequisites:
   • .NET 8 SDK
   • Node.js 20+
2. Backend setup:
   • Restore dependencies with dotnet build.
   • Configure appsettings.* or user secrets with LLM provider keys and defaults.
3. Frontend setup:
   • Use Node.js 20.19+ or 22.12+ (Vite 7 requirement).
   • cd frontend && npm install
   • Copy .env.example to .env.local (optional) and set VITE_API_BASE_URL.
   • npm run dev to start the UI on port 5173.
4. Launch both projects and visit the frontend to start chatting with your database.

Detailed backend/frontend instructions will be added as the implementation progresses.

Roadmap

• Bootstrap backend services with Semantic Kernel integration.
• Implement MongoDB tool plugins and query safety checks.
• Build chat orchestration and streaming response handling.
• Create frontend chat UI with chart/table rendering.
• Document deployment options (Azure App Service, Vercel, etc.).

Contributing

Suggestions and pull requests are welcome. Please open an issue to discuss major changes before submitting a PR.

License

MIT