Developer Chat Platform with AI Integration

A demo-ready real-time chat platform for developers. Users can register, start one-to-one chats with other developers, exchange live messages over WebSockets, and call the AI assistant inside the same shared thread with an @ai message.

The current implementation is Node.js/Express on the backend and React/Vite on the frontend. The project prompt mentions Python, but this repository contains a JavaScript backend, so the audit refined the stack already present instead of rewriting it.

Key Features

• Developer-to-developer chat creation with participant assignment
• Dashboard chat list scoped to the logged-in user
• Full persisted chat history on thread open
• Socket.IO real-time messaging with one room per chat
• @ai trigger inside normal chats, with AI replies broadcast to both users
• MongoDB persistence for users, chats, and messages
• JWT authentication with HTTP-only cookie support and bearer token fallback
• Optional Redis logout token blacklist
• Gemini integration with local fallback responses for credential-free demos

Tech Stack

• Frontend: React, Vite, Tailwind CSS, Axios, Socket.IO Client
• Backend: Node.js, Express, Socket.IO, MongoDB, Mongoose
• Auth: bcryptjs, JSON Web Tokens, cookies
• AI: Gemini API, local fallback when no API key is configured
• Optional: Redis for logout token blacklisting

Setup

Install backend dependencies:

cd backend
npm install

Install frontend dependencies:

cd ../frontend
npm install

Create local environment files:

copy backend\.env.example backend\.env
copy frontend\.env.example frontend\.env

Start MongoDB locally, then run the backend:

cd backend
npm run dev

Run the frontend in another terminal:

cd frontend
npm run dev

Open http://localhost:5173.

Environment Variables

Backend backend/.env:

PORT=3000
CLIENT_URL=http://localhost:5173
MONGODB_URI=mongodb://127.0.0.1:27017/developer_chat_ai
MONGODB_SERVER_SELECTION_TIMEOUT_MS=5000
JWT_SECRET=replace-with-a-long-random-secret
JWT_EXPIRES_IN=24h
GEMINI_API_KEY=YOUR_API_KEY
GEMINI_MODEL=gemini-2.5-flash
REDIS_URL=

Frontend frontend/.env:

VITE_API_URL=http://localhost:3000

GEMINI_API_KEY is optional for demos. If it is missing or set to YOUR_API_KEY, the backend returns a deterministic local AI response.

How Real-Time Chat Works

1. The frontend authenticates with the JWT and opens a Socket.IO connection.
2. The backend validates the socket token and joins the user to a private user:<id> room.
3. When a user opens a chat, the frontend emits chat:join; the backend verifies membership and joins chat:<chatId>.
4. New messages are saved to MongoDB first, then broadcast to the chat room as message:new.
5. Chat list summaries are updated with chat:created and chat:updated events.
6. If a message starts with @ai, the backend generates an AI reply, saves it as a normal chat message, and broadcasts it to the same room.

Demo Flow

1. Register two users in separate browsers or profiles.
2. User A selects User B from the developer picker and creates a chat.
3. User B sees the chat appear on their dashboard after login, or instantly if already online.
4. Either user opens the chat and sends a message.
5. Both users see the message live without refreshing.
6. Send @ai explain this code to add an AI response visible to both users.

Useful Scripts

Backend:

npm start
npm run dev
npm run check

Frontend:

npm run dev
npm run build
npm run lint
npm run preview

Future Scope

• Add WebContainer-powered code execution rooms for collaborative coding
• Add typing indicators and read receipts
• Add group chats and role-based project channels
• Add file/code snippet attachments
• Add automated integration tests for Socket.IO events
• Add production Socket.IO scaling with a Redis adapter