Meeting AI: Full-Stack LLM-Powered Meeting Management Platform

Live Demo

Transforming Conversations into Actionable Intelligence

Meeting AI is a robust, production-grade full-stack application engineered to eliminate the inefficiencies of traditional meeting workflows. By integrating high-speed Large Language Models (LLMs) and real-time communication technologies, it automates the entire post-meeting process—from transcription to task assignment—allowing teams to focus purely on collaboration and decision-making.

This platform is built for modern teams seeking a seamless, secure, and highly performant solution for meeting documentation and follow-up.

Core Capabilities and Value Proposition

Feature	Technical Implementation	Key Benefit
Real-Time Conferencing	WebRTC for peer-to-peer audio/video streams, managed by Socket.io signaling.	Conduct secure, high-quality live meetings without relying on external services.
High-Fidelity Transcription	Utilizes the OpenAI Whisper API (whisper-1) for accurate audio-to-text conversion, including automatic translation to English.	Provides a complete, corrected, and reliable record of every discussion point.
Intelligent Summarization	GPT-4o mini generates structured meeting notes, summarizing key points, decisions, and next steps in a professional format.	Reduces manual effort and ensures consistent, high-quality documentation for every meeting.
Automated Task Extraction	LLM-driven analysis of the transcript to identify action items and assign them to the correct participant based on conversational context.	Drives accountability and ensures critical follow-up tasks are never missed.
Contextual Q&A Chatbot	An integrated AI assistant that uses the meeting transcript as its sole knowledge base to answer specific user queries.	Enables quick information retrieval and deep understanding of past discussions.
Secure Organization Management	Supabase Auth and PostgreSQL RLS enforce domain-based team access and granular data security policies.	Guarantees a private and compliant workspace for sensitive corporate data.

Technical Architecture Deep Dive

Meeting AI is structured as a modular, high-performance application, separating the user interface from the business logic and AI processing.

Frontend: React, Vite, and Tailwind CSS

• Framework: React.js with Vite for a blazing-fast development and build experience.
• Styling: Tailwind CSS implements a professional, responsive UI with a modern "glassmorphism" aesthetic.
• Real-Time: Custom React Hooks manage the complex state of WebRTC connections and Socket.io events.

Backend: Node.js, Express, and OpenAI

• API Server: Node.js with Express provides a scalable RESTful API for managing users, meetings, and chat history.
• AI Integration: Direct use of the OpenAI SDK for transcription (whisper-1) and chat completions (gpt-4o-mini) — notes, task extraction, and Q&A.
• Data & Storage: Supabase serves as the unified backend for:
  • Authentication: User sign-up and sign-in.
  • Database: PostgreSQL with robust Row Level Security (RLS) policies for data isolation.
  • Storage: Secure storage of raw audio files.

AI Stack

Meeting AI uses the OpenAI API for all LLM operations:

Task	Model
Audio transcription & translation	whisper-1
Grammar correction, notes, tasks, chat	gpt-4o-mini

Configure OPENAI_API_KEY in backend/.env. Audio filenames must use a supported extension (.mp3, .m4a, .wav, .webm, etc.) so Whisper accepts the upload.

Getting Started (Developer Setup)

Prerequisites

• Node.js 16+ and npm
• A Supabase account and project
• An OpenAI API Key

1. Supabase Configuration

1. Create a new Supabase project.
2. Execute the SQL schema from supabase_schema.sql to set up the meetings, tasks, and chat_messages tables, including all necessary RLS policies.
3. Create a public storage bucket named meeting-audio.

2. Environment Variables

Configure your credentials in the respective .env files:

**backend/.env**

PORT=5000
SUPABASE_URL=your_supabase_project_url
SUPABASE_SERVICE_KEY=your_supabase_service_role_key
OPENAI_API_KEY=your_openai_api_key
# Optional: SMTP details for email invitations

**frontend/.env**

VITE_SUPABASE_URL=your_supabase_project_url
VITE_SUPABASE_ANON_KEY=your_supabase_anon_key
VITE_API_URL=http://localhost:5000/api

3. Installation and Local Run

# Install dependencies
cd backend && npm install
cd ../frontend && npm install

# Run the application
# Terminal 1: Backend Server
cd backend
npm run dev

# Terminal 2: Frontend Client
cd frontend
npm run dev

The application will be accessible at http://localhost:5173.

Mobile app (Flutter)

A Flutter iOS/Android client lives in meeting_ai/. See meeting_ai/README.md for setup, backend URL configuration, and build instructions.

📄 License

This project is released under the MIT License.