Dadi - AI Assistant Chatbot

A secure, Azure-hosted AI chatbot application with Google OAuth authentication and whitelist-based access control.

Features

• Google OAuth Authentication - Secure login via Azure Easy Auth
• Whitelist-based Access - Control who can use the application
• Azure OpenAI Integration - Powered by GPT models
• Chat Modes - Switch between fun and normal personality modes (per-user, persisted)
• Conversation History - Browse, pin, and delete past conversations (stored in-browser)
• Slash Commands - Built-in tool framework with /idea, /ideas, /issue, /chatmode
• Idea Bank - Submit and list ideas stored in Azure Blob Storage
• GitHub Issue Creation - Create GitHub issues directly from chat with /issue
• AI Auto-Resolve - Automatically resolves GitHub issues by generating code and opening PRs
• Mobile-Responsive UI - Carousel-style panel navigation on mobile devices
• PWA Support - Installable as a Progressive Web App via service worker
• Modern Chat Interface - Clean, responsive UI with markdown rendering and syntax highlighting
• Fully Cloud-hosted - Azure App Service
• Auto-deployment - GitHub Actions CI/CD pipeline

Architecture

• Frontend: TypeScript + Vite + Vanilla TS → bundled into wwwroot/, served by the backend
• Backend: C# ASP.NET Core (.NET 10) → Azure App Service (app-dadi)
• Storage: Azure Blob Storage (ideas, user preferences)
• AI: Azure OpenAI Service
• Auth: Azure Easy Auth on App Service (Google)
• GitHub Integration: GitHub API (issue creation, auto-resolve PRs)
• Deployment: Single GitHub Actions workflow (builds frontend + backend together)

The frontend and backend are deployed together as a single Azure App Service. The frontend is built by CI and copied into backend-csharp/wwwroot/ before the .NET app is published. The ASP.NET Core app serves static files from wwwroot/ and falls back to index.html for SPA routing.

Quick Start

Prerequisites

• Azure subscription with:
  • Azure App Service (Windows or Linux, .NET 10)
  • Azure Storage account (for ideas and user preferences)
  • Azure OpenAI resource with deployed model
  • Easy Auth configured on App Service (Google OAuth)
• GitHub repository with required secrets configured

Setup

1. Configure Azure Resources

   • Create an Azure App Service targeting .NET 10
   • Create an Azure Storage account for blob data
   • Create an Azure OpenAI resource and deploy a model

2. Configure App Service

   • Set environment variables in Azure App Service → Configuration → Application settings:
     • AZURE_OPENAI_ENDPOINT
     • AZURE_OPENAI_KEY
     • AZURE_OPENAI_DEPLOYMENT
     • WHITELISTED_EMAILS
     • AZURE_STORAGE_CONNECTION_STRING
     • GITHUB_TOKEN (GitHub personal access token for issue creation and auto-resolve)

3. Configure Easy Auth

   • Enable Authentication on App Service
   • Add Google identity provider
   • Configure redirect URLs

4. Deploy

   • Push to master branch to trigger automatic deployment via GitHub Actions
   • The workflow builds the frontend, copies it to wwwroot/, then publishes the .NET project to Azure App Service

5. Access

   • Navigate to https://app-dadi.azurewebsites.net
   • Login with Google using a whitelisted email
   • Start chatting with Dadi!

Documentation

• Azure Setup Guide - Create Azure resources
• Function App Setup - Configure the App Service
• Configuration Guide - Configure the application
• GitHub Secrets Setup - Configure GitHub Actions

Project Structure

dadi/
├── frontend/              # Frontend application (Vanilla TS)
│   ├── src/
│   │   ├── components/    # UI components (Header, Chat, Message, ConversationSidebar, InfoPanel, ConfirmDialog, LoginScreen)
│   │   ├── services/      # API services (chat, user, auth, chatmode, idea, conversationStorage, conversationStore)
│   │   ├── styles/        # CSS stylesheets
│   │   ├── tools/         # Slash command tools (idea, issue, chatmode, toolRegistry)
│   │   ├── types/         # TypeScript type definitions
│   │   └── utils/         # Utility functions (authGuard, deviceMode, debugMode, errors, tokenUtils)
│   └── public/            # Static assets
├── backend-csharp/        # C# ASP.NET Core backend
│   ├── Controllers/       # HTTP controllers (User, Chat, Ideas, Issues, UserPreferences)
│   ├── Models/            # Data models
│   ├── Services/          # Business logic (Auth, OpenAI, IdeaStorage, UserPreferences, GitHub, AutoResolve)
│   └── wwwroot/           # Built frontend assets (populated by CI)
├── persona/               # AI persona definitions
├── docs/                  # Documentation
└── .github/workflows/     # GitHub Actions CI/CD

Development

Frontend

cd frontend
npm install
npm run dev       # Start development server (port 5173)
npm run build     # Build for production

Backend (C#)

cd backend-csharp
dotnet restore
dotnet build
dotnet run        # Start ASP.NET Core locally (port 5000/5001 or as configured)

Environment Variables

Frontend (.env.development):

VITE_FUNCTION_APP_URL=http://localhost:5000
VITE_BYPASS_AUTH_FOR_LOCAL_DEV=true

Backend (appsettings.Development.json):

{
  "AZURE_STORAGE_CONNECTION_STRING": "UseDevelopmentStorage=true",
  "AZURE_OPENAI_ENDPOINT": "https://your-openai.openai.azure.com/",
  "AZURE_OPENAI_KEY": "your-key",
  "AZURE_OPENAI_DEPLOYMENT": "chat",
  "WHITELISTED_EMAILS": "user1@example.com,user2@example.com",
  "BYPASS_AUTH_FOR_LOCAL_DEV": "true",
  "GITHUB_TOKEN": "your-github-pat-here"
}

Slash Commands

Command	Description
/idea <text>	Submit an idea (max 500 chars)
/idea delete <n>	Delete idea number n from the list
/ideas	List all submitted ideas
/issue <title>	Create a GitHub issue (max 256 chars)
/chatmode	Show current chat mode
/chatmode fun|normal	Change chat mode

Auto-Resolve

The AutoResolveService can be triggered to resolve open GitHub issues. It:

1. Fetches open GitHub issues labelled from-chat that do not yet have the autoresolve label
2. Uses Azure OpenAI to generate an implementation plan and code changes
3. Commits the changes to a new branch and opens a pull request that closes the issue

Requires GITHUB_TOKEN with repo scope and AZURE_OPENAI_* variables to be configured.

Security

• Authentication required for all endpoints
• Email whitelist enforced on backend
• Easy Auth handles OAuth flow
• API keys stored in Azure App Service configuration (not in code)
• HTTPS enforced by Azure
• CORS configured for local development only

License

ISC

Author

Built with ❤️ and GitHub Copilot