MeetMind - AI-Powered Meeting Assistant

A full-stack web application for AI-powered meeting assistance using the MERN stack, with real-time transcription, intelligent meeting summaries and monthly knowledge base views powered by modern speech-to-text and language models.

Features

• 🎙️ Real-time Transcription: Transcribe meetings in real-time using Deepgram.
• 📝 AI-Powered Summaries: Automatically generate concise summaries and meeting minutes.
• 📚 Knowledge Base: Store past meetings in MongoDB Atlas and search them using Atlas Vector Search for semantic recall.
• 📅 Monthly View: Showing all meetings from a selected month with detailed analytics.
• 🌍 Multi-language Support: Support multi-language transcription and cultural context adaptation.

Technology Stack

Frontend

• React (with Vite)
• React Router
• Socket.IO Client
• Tailwind CSS
• Axios
• React Hot Toast

Backend

• Node.js + Express
• Socket.IO (WebSocket handling)
• MongoDB + Mongoose
• MongoDB Atlas Vector Search
• Deepgram SDK (Speech-to-text)
• OpenAI API (GPT for summarization and task extraction)
• Nodemailer (Email invitations)
• JWT Authentication

Prerequisites

• Node.js (v18 or higher)
• MongoDB Atlas account (with Vector Search enabled)
• Deepgram API key
• OpenAI API key
• Email service credentials (SMTP for email invites)

Installation

Backend Setup

1. Navigate to the backend directory:

cd backend

2. Install dependencies:

npm install

3. Create a .env file in the backend directory:

PORT=5001
NODE_ENV=development
FRONTEND_URL=http://localhost:5173

MONGODB_URI=your_mongodb_atlas_connection_string
JWT_SECRET=your_jwt_secret_here
DEEPGRAM_API_KEY=your_deepgram_api_key
OPENAI_API_KEY=your_openai_api_key
EMAIL_USER=your_gmail@gmail.com
EMAIL_PASSWORD=your_16_char_app_password

1. MONGODB_URI=your_mongodb_atlas_connection_string
Create a free MongoDB Atlas cluster, create a database user, then click Connect → Drivers → copy the mongodb+srv:// connection string and paste it here. Replace <password> with your DB user password.

2. JWT_SECRET=your_jwt_secret_here
Any strong random string, e.g. generated via an online UUID generator or openssl rand -hex 32.

3. DEEPGRAM_API_KEY=your_deepgram_api_key
Create a Deepgram account → go to the API Keys section → click Create API Key → copy the key and paste it here.

4. OPENAI_API_KEY=your_openai_api_key
Go to https://platform.openai.com → sign in / sign up → open the API Keys page → click Create new secret key → copy the key once and paste it here.

5. EMAIL_PASSWORD=your_16_char_app_password
→ Enable 2FA: Google Account → Security → 2-Step Verification
→ App Passwords → Mail → "MeetMind" → Generate
→ Copy 16-char password (remove spaces)
→ Set EMAIL_PASSWORD=abcdefghijklmnop in .env

NOTE: This project uses your own OpenAI account so if the free or paid quota is exhausted, in that case you can either:
i. Create a new OpenAI account, generate a fresh API key, and update OPENAI_API_KEY in .env, OR
ii. Add credit / increase quota in your existing OpenAI account.

4. Start the backend server:

npm start

Frontend Setup

1. Navigate to the frontend directory:

cd frontend

2. Install dependencies:

npm install

3. Create a .env file in the frontend directory:

VITE_API_URL=http://localhost:5001/api
VITE_SOCKET_URL=http://localhost:5001
DEEPGRAM_API_KEY=your_deepgram_api_key

4. Start the frontend development server:

npm run dev

Quick Start

1. git clone https://github.com/akansha-02/MeetMind.git
2. cd MeetMind/backend && npm install
3. cd MeetMind/frontend && npm install
4. create backend/.env and frontend/.env ->npm start (backend) | npm run dev (frontend) ->open http://localhost:5173

Usage

AI features (transcription, summaries, knowledge base search) require valid Deepgram and OpenAI keys configured in backend/.env as described above.

1. Start both backend and frontend servers
2. Open your browser and navigate to http://localhost:5173
3. Register a new account or login
4. Create a new meeting
5. Start recording or upload an audio file
6. View real-time transcription
7. Complete the meeting to generate AI summaries
8. View monthly knowledge base → Filter meetings by month
9. Search the knowledge base for context

API Endpoints

Authentication

• POST /api/auth/register - Register a new user
• POST /api/auth/login - Login user
• GET /api/auth/me - Get current user (protected)

Meetings

• GET /api/meetings - Get all meetings (protected)
• GET /api/meetings/:id - Get single meeting (protected)
• POST /api/meetings - Create meeting (protected)
• PUT /api/meetings/:id - Update meeting (protected)
• DELETE /api/meetings/:id - Delete meeting (protected)
• POST /api/meetings/:id/complete - Complete meeting and generate AI content (protected)
• POST /api/meetings/:id/process - Process meeting with AI (protected)
• POST /api/meetings/:id/invite - Send email invites (protected)
• GET /api/meetings/monthly/:year/:month - Monthly meetings view (protected)

Transcripts

• POST /api/transcripts/upload - Upload and transcribe audio file (protected)
• GET /api/transcripts/:meetingId - Get transcript for meeting (protected)

Knowledge Base

• POST /api/knowledge-base/search - Search knowledge base (protected)
• GET /api/knowledge-base/monthly/:year/:month - Monthly knowledge base view

Socket.IO Events

Client to Server

• join-meeting - Join a meeting room
• start-transcription - Start live transcription
• stop-transcription - Stop live transcription
• audio-data - Send audio data chunks

Server to Client

• joined-meeting - Confirmation of joining meeting
• transcription-started - Transcription has started
• transcription-stopped - Transcription has stopped
• transcript-interim - Interim transcription results
• transcript-final - Final transcription results
• transcription-error - Transcription error occurred

Project Structure

MeetMind/
├── backend/
│   ├── src/
│   │   ├── config/         # Configuration files
│   │   ├── controllers/    # Route controllers
│   │   ├── models/         # Mongoose models
│   │   ├── routes/         # Express routes
│   │   ├── services/       # Business logic services
│   │   ├── middleware/     # Express middleware
│   │   ├── socket/         # Socket.IO handlers
│   │   ├── utils/          # Utility functions
│   │   └── server.js       # Server entry point
│   └── package.json
├── frontend/
│   ├── src/
│   │   ├── components/     # React components
│   │   ├── pages/          # Page components
│   │   ├── hooks/          # Custom React hooks
│   │   ├── services/       # API and Socket services
│   │   ├── contexts/       # React contexts
│   │   ├── utils/          # Utility functions
│   │   ├── App.jsx         # Main app component
│   │   └── main.jsx        # Entry point
│   └── package.json
└── README.md

Development

• Backend runs on http://localhost:5001
• Frontend runs on http://localhost:5173

Troubleshooting

• Backend not starting or login/register failing:
  • Ensure MONGODB_URI and JWT_SECRET are set in backend/.env.
  • AI keys (OPENAI_API_KEY, DEEPGRAM_API_KEY, etc.) are optional; without them, AI features are disabled but auth works.
  • Check the backend health: GET http://localhost:5001/api/health.
  • If port 5001 is in use, free it on Windows:

    Get-NetTCPConnection -LocalPort 5001 | Select-Object OwningProcess
    Stop-Process -Id <PID> -Force

• Frontend cannot reach backend:
  • Set VITE_API_URL in frontend/.env to http://localhost:5001/api.
  • Verify CORS origin matches FRONTEND_URL in backend .env.

License

ISC

Contributing

Contributions are welcome! Please feel free to submit a Pull Request. For major changes, please open an issue first to discuss what you would like to change.