Skip to content

hasancoded/ai-meeting-copilot

BranchesTags

Repository files navigation

AI Meeting Copilot

License: MIT Node.js TypeScript Tests

AI-powered meeting assistant that transcribes audio and video recordings, generates executive summaries, extracts action items, and tracks decisions. Users bring their own API key — no shared credentials, no hidden costs.

Features

  • Audio Transcription — Supports MP3, WAV, M4A, MP4, and WebM via OpenAI Whisper or Google Gemini
  • AI Summarization — Concise executive summaries from full meeting transcripts
  • Action Item Extraction — Automatically identifies tasks, owners, and due dates
  • Decision Tracking — Captures key decisions made during meetings
  • Per-User API Keys — Users supply and manage their own Gemini or OpenAI key via the Settings page; keys are encrypted at rest with AES-256-GCM
  • Model Selection — Users choose their preferred model (Gemini 2.5 Flash, GPT-4o-mini, etc.) from the Settings page
  • Secure Authentication — JWT tokens in httpOnly cookies; all data is user-isolated
  • Modern UI — Responsive React interface with drag-and-drop file upload

Architecture

graph TB
    subgraph Client["Client Layer"]
        UI["React Frontend<br/>Vite + TailwindCSS"]
    end

    subgraph API["API Layer"]
        REST["Express REST API<br/>JWT Auth + CORS"]
        AuthMW["Auth Middleware<br/>JWT Validation"]
        RouteH["Route Handlers<br/>Zod Validation"]
    end

    subgraph Service["Service Layer"]
        AISvc["AI Service<br/>Provider Pattern"]
        TransSvc["Transcription Service<br/>Provider Pattern"]
        FileSvc["File Handler<br/>Multer + Cleanup"]
        KeySvc["API Key Service<br/>AES-256-GCM"]
    end

    subgraph Data["Data Layer"]
        ORM["Prisma ORM<br/>Type-Safe Queries"]
        Database[("Database<br/>SQLite / PostgreSQL")]
    end

    subgraph External["External Services"]
        OpenAI["OpenAI API<br/>GPT-4o + Whisper"]
        Gemini["Google Gemini<br/>2.5 Flash"]
    end

    UI -->|HTTP/JSON| REST
    REST --> AuthMW
    AuthMW --> RouteH
    RouteH --> AISvc
    RouteH --> TransSvc
    RouteH --> FileSvc
    RouteH --> KeySvc
    RouteH --> ORM

    AISvc -.->|API| OpenAI
    AISvc -.->|API| Gemini
    TransSvc -.->|API| OpenAI
    TransSvc -.->|API| Gemini

    ORM -->|SQL| Database
    FileSvc -->|I/O| Storage["File System<br/>uploads/"]
Loading

Key Design Patterns

  • Provider Pattern — Pluggable AI and transcription services with a stub implementation for zero-config development
  • Per-User Key Injection — API keys are decrypted at request time and injected into provider constructors; the raw key is never logged or stored in plaintext
  • Repository Pattern — Prisma ORM abstracts all database access with full type safety
  • Retry Logic — Exponential backoff on all external API calls

Tech Stack

Backend

Concern Technology
Runtime Node.js 20
Framework Express.js
Language TypeScript 5.4
Database ORM Prisma (SQLite / PostgreSQL)
Authentication JWT + bcrypt
Validation Zod
File Upload Multer
Encryption Node.js crypto (AES-256-GCM)
Testing Jest + Supertest

Frontend

Concern Technology
Framework React 19
Build Tool Vite
Styling TailwindCSS
Routing React Router 6
State Zustand
HTTP Client Axios
Icons Lucide React

Infrastructure

  • Docker + Docker Compose for production deployment
  • PostgreSQL 16 for production database
  • Multi-stage Dockerfile for minimal image size

Project Structure

ai-meeting-copilot/
├── docs/
│   ├── API.md                   # Complete API reference
│   └── DETAILED_GUIDE.md        # Deployment and technical guide
├── server/
│   ├── prisma/
│   │   ├── migrations/          # Prisma migration history
│   │   └── schema.prisma        # Database schema
│   └── src/
│       ├── middleware/
│       │   └── auth.ts          # JWT authentication middleware
│       ├── routes/
│       │   ├── apiKey.ts        # API key management (save, delete, status)
│       │   ├── auth.ts          # Register, login, logout
│       │   ├── health.ts        # Health check
│       │   └── meetings.ts      # Meeting CRUD, upload, process
│       ├── services/
│       │   ├── ai/
│       │   │   ├── gemini.ts    # Google Gemini provider
│       │   │   ├── openai.ts    # OpenAI provider
│       │   │   ├── provider.ts  # AIProvider interface
│       │   │   └── stub.ts      # Stub for development
│       │   └── transcription/
│       │       ├── gemini-transcriber.ts
│       │       ├── provider.ts  # TranscriptionProvider interface
│       │       ├── stub.ts
│       │       └── whisper.ts
│       ├── tests/
│       │   ├── apiKey.test.ts
│       │   ├── auth.test.ts
│       │   ├── meetings.test.ts
│       │   └── setup.ts
│       ├── utils/
│       │   ├── encryption.ts    # AES-256-GCM encrypt/decrypt
│       │   ├── jwt.ts
│       │   └── parsing.ts
│       ├── db.ts                # Prisma client singleton
│       ├── env.ts               # Validated environment variables (Zod)
│       ├── index.ts             # Express app entry point
│       └── types.ts             # Shared TypeScript types
├── web/
│   └── src/
│       ├── components/
│       │   ├── ui/              # Button, Badge, Card, Spinner, etc.
│       │   ├── CreateMeetingModal.tsx
│       │   ├── DeleteMeetingModal.tsx
│       │   ├── FileDrop.tsx
│       │   ├── Layout.tsx
│       │   └── ProtectedRoute.tsx
│       ├── lib/
│       │   └── api.ts           # Axios API client
│       ├── pages/
│       │   ├── Dashboard.tsx
│       │   ├── Login.tsx
│       │   ├── MeetingDetail.tsx
│       │   ├── Register.tsx
│       │   └── Settings.tsx     # API key and model configuration
│       ├── store/
│       │   └── auth.ts          # Zustand auth store
│       ├── App.tsx
│       ├── index.css
│       └── main.tsx
├── .dockerignore
├── .gitignore
├── CONTRIBUTING.md
├── docker-compose.yml
├── Dockerfile
├── LICENSE
├── package.json                 # Workspace root
└── README.md

Quick Start

Prerequisites

  • Node.js 20 or higher
  • npm

1. Clone and install

git clone https://github.com/hasancoded/ai-meeting-copilot.git
cd ai-meeting-copilot
cd server && npm install && cd ../web && npm install && cd ..

2. Configure the backend

cp server/.env.development.example server/.env.development

The minimum required values in server/.env.development:

PORT=4000
NODE_ENV=development
FRONTEND_URL=http://localhost:5173
DATABASE_URL="file:./dev.db"

# Generate with: node -e "console.log(require('crypto').randomBytes(32).toString('hex'))"
JWT_SECRET=your_jwt_secret_min_32_chars
ENCRYPTION_SECRET=your_encryption_secret_min_32_chars

# Use stub for development — no API key required
AI_PROVIDER=stub
TRANSCRIBE_PROVIDER=stub

3. Configure the frontend

cp web/.env.example web/.env
# VITE_API_URL is already set to http://localhost:4000

4. Run database migrations

cd server
DATABASE_URL="file:./dev.db" npx prisma migrate dev

5. Start the servers

# Terminal 1 — backend
cd server && npm run dev

# Terminal 2 — frontend
cd web && npm run dev

Open http://localhost:5173.

Usage

  1. Register — Create an account with email and password
  2. Configure API Key — Go to Settings, choose a provider (Gemini or OpenAI) and model, paste your API key
  3. Create a Meeting — Click "New Meeting" and enter a title
  4. Upload Audio — Drag and drop or select an audio/video file (max 100 MB)
  5. Process Meeting — Click "Process Meeting" to transcribe and analyze
  6. View Results — Switch between tabs for Summary, Transcript, Action Items, and Decisions

For development without an API key, leave AI_PROVIDER=stub and TRANSCRIBE_PROVIDER=stub. The stub provider returns realistic-looking synthetic data so the full UI can be exercised without incurring API costs.

API Endpoints

Full documentation: docs/API.md

Authentication

Method Path Description
POST /api/auth/register Create account
POST /api/auth/login Authenticate
POST /api/auth/logout End session

Meetings

Method Path Description
GET /api/meetings List user's meetings
POST /api/meetings Create meeting
GET /api/meetings/:id Get meeting details
POST /api/meetings/:id/upload Upload audio file
POST /api/meetings/:id/process Transcribe and analyze
DELETE /api/meetings/:id Delete meeting

User Settings

Method Path Description
GET /api/user/api-key/status Check API key status (never returns the key)
PUT /api/user/api-key Save or update API key, provider, and model
DELETE /api/user/api-key Remove saved API key

Health

Method Path Description
GET /api/health API status check

Environment Variables

Server (server/.env.development or server/.env.production)

Variable Required Description
JWT_SECRET Yes JWT signing secret — minimum 32 characters
ENCRYPTION_SECRET Yes AES-256-GCM key for API key encryption — minimum 32 characters
DATABASE_URL Yes Prisma database connection string
PORT No Server port (default: 4000)
NODE_ENV No development or production
FRONTEND_URL No CORS allowed origin (default: http://localhost:5173)
AI_PROVIDER No stub, gemini, or openai (default: stub)
TRANSCRIBE_PROVIDER No stub, gemini, or whisper (default: stub)

GEMINI_API_KEY and OPENAI_API_KEY are not required at the server level. Each user supplies their own key through the Settings page. The server encrypts it with ENCRYPTION_SECRET before storing it in the database.

Frontend (web/.env)

Variable Required Description
VITE_API_URL Yes Backend API base URL (default: http://localhost:4000)

Testing

cd server
npm test

48 tests covering authentication, meeting CRUD, file upload, processing pipeline, API key management, and error handling. Tests run against a dedicated in-memory SQLite database and use stub AI/transcription providers — no real API keys required.

Production Deployment

Docker Compose

# Copy and edit production environment
cp server/.env.production.example server/.env.production
# Edit: set JWT_SECRET, ENCRYPTION_SECRET, DATABASE_URL, FRONTEND_URL
# Leave AI_PROVIDER=gemini and TRANSCRIBE_PROVIDER=gemini (users supply their own keys)

docker-compose up -d

Manual checklist before going live

  • JWT_SECRET — unique, 64+ character hex value
  • ENCRYPTION_SECRET — unique, 64+ character hex value (if rotated, existing users must re-enter their key)
  • DATABASE_URL — PostgreSQL connection string
  • FRONTEND_URL — production frontend domain
  • AI_PROVIDER and TRANSCRIBE_PROVIDER set to gemini or openai

Contributing

Contributions are welcome. Please read CONTRIBUTING.md before opening a pull request.

License

MIT — see LICENSE.

About

AI-powered meeting assistant that transcribes audio/video files, generates summaries, extracts action items, and builds a searchable knowledge base. Built with Node.js, Express, React, and Google Gemini AI.

Topics

react nodejs docker productivity express typescript ai rest-api postgresql saas full-stack openai summarization jwt-authentication transcription whisper prisma tailwindcss meeting-assistant gemini-ai

Resources

Readme

License

MIT license

Contributing

Contributing
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages