Meet AI - AI-Powered Video Meeting SaaS

---

🚀 A SaaS platform that enables AI-powered video meetings with real-time AI agents, meeting transcripts, summaries, playback, and AI Q&A.
Built with React, Express, Passport.js, Drizzle ORM, and Stream.io.

---

📑 Table of Contents

• About
• Features
• Getting Started
• Running Tests
• Usage
• Deployment
• Built Using
• License

---

🧐 About

Meet AI is a full-stack SaaS platform designed for hosting video meetings enhanced with AI-driven agents (e.g., tutors, coaches, assistants). These agents interact in real-time during meetings, providing dynamic support. Post-meeting, the platform generates summaries, searchable transcripts, call recordings, and enables AI-powered Q&A for reviewing discussions.

This project serves as a scalable SaaS template for building AI-enhanced video collaboration tools.

---

✨ Features

• 🔐 User Authentication: Supports Google OAuth, and Email login via Passport.js
• 📞 Real-time Video Calls: Powered by Stream.io Video SDK
• 💬 Live In-call Chat: Integrated with Stream.io Chat SDK
• 🤖 AI Meeting Assistants: Context-aware, real-time responses using OpenAI
• 📝 Post-Meeting Intelligence:
  • AI-generated structured summaries
  • Searchable transcripts with timestamps
  • Recording playback
  • Chat-style AI follow-ups for meeting insights
• 📂 Dashboard: Manage upcoming, live, and past meetings
• 📱 Responsive UI: Built with React and TailwindCSS
• 💳 Subscription-Ready Architecture: Express backend with Stripe integration support

---

🏁 Getting Started

Prerequisites

You’ll need:

• Node.js >= 20
• pnpm (preferred package manager)
• PostgreSQL database (recommended: Neon.tech)
• Stream.io account for Video and Chat SDK
• OpenAI account for AI agents

Installation

1. Clone the repository

    git clone https://github.com/myselfsandip/meet-ai.git
    cd meet-ai

2. Install dependencies

pnpm install

3. Set up environment variables

Create a .env file in the project root

Database

DATABASE_URL=postgres://user:password@host/db

Authentication

SESSION_SECRET=your_secret_key
GOOGLE_CLIENT_ID=your_google_id
GOOGLE_CLIENT_SECRET=your_google_secret

Stream.io

STREAM_KEY=xxx
STREAM_SECRET=xxx

OpenAI

OPENAI_API_KEY=sk-xxxx

Set up the database with Drizzle ORM

pnpm drizzle-kit push

Run development servers

Start the backend API (Express):

pnpm dev:server

Start the frontend (React client):

pnpm dev:client

🔧 Running the Tests

Run automated tests with:

pnpm test

Types of Tests

Unit Tests: Cover Drizzle models and utility functions Integration Tests: Validate API endpoints, authentication, and meeting creation

🎈 Usage

Register or log in using Google, or Email via Passport.js. Create an AI Agent by defining its behavior or persona. Start a new meeting and select your AI agent. Engage in real-time conversations with the AI and other participants. After the meeting, access summaries, transcripts, and recordings. Use the AI assistant to ask questions about the meeting content.

🚀 Deployment

• Frontend (React): Deploy to AWS Cloudfront CDN via S3 Bucket .

• Backend (Express API): Deploy to AWS EC2 VM.

• Database: Use Neon.tech for Remote Postgres Database

⛏️ Built Using

• React.js - Frontend UI
• TailwindCSS - Styling
• SHADCN UI - UI Components
• Tanstack Query - API Data Fetching & Caching
• Zustand - State Management
• Express.js - Backend API
• Passport.js - Authentication
• Drizzle ORM - Database ORM (Postgres)
• Stream.io - Video and Chat SDK
• OpenAI - AI-powered meeting features

License

Licensed under the MIT license.

You're free to use my code! Just make sure to remove all my personal information before publishing your website. It's awesome to see my code being useful to someone!