reply.

A fully automated system that monitors your YouTube videos 24/7 and replies to comments based on keywords with human-like behavior.

Live at: tryreply.app

What It Does

• Monitors your YouTube channel videos automatically
• Filters comments by custom keywords (e.g., "MCP", "help", "tutorial")
• Replies automatically with customizable templates
• Prevents duplicate replies with sub-millisecond database checks
• Uses human-like delays to avoid YouTube ban
• Tracks quota usage (10,000 units/day limit)
• Provides analytics dashboard with real-time system health

Architecture

+------------------+
|   Next.js        |  Frontend (Vercel)
|   Frontend       |  - Video management
|   (Vercel)       |  - Settings UI
+--------+---------+  - Analytics
         |
         | HTTPS/REST
         v
+------------------+
|   FastAPI        |  Backend (AWS EC2)
|   Backend        |  - OAuth authentication
|   (Docker)       |  - YouTube API integration
+--------+---------+  - Auto-reply engine
         |
    +----+----+----------+----------+
    |         |          |          |
    v         v          v          v
+--------+ +------+ +--------+ +---------+
|  RDS   | |Redis | |YouTube | | Celery  |
|Postgres| |(Docker)| |  API   | | Worker |
+--------+ +------+ +--------+ +---------+

Infrastructure

Service	Spec	Cost/Month
AWS EC2	t3.small (Ubuntu)	~$15
AWS RDS	PostgreSQL db.t3.micro	~$13
Elastic IP	1 (attached)	FREE
Vercel	Hobby tier	FREE
TOTAL		~$28/month

Quick Start

1. Clone Repository

git clone https://github.com/farizanjum/reply.git
cd reply-comments

2. Backend Setup (Local)

cd backend

# Create virtual environment
python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Set up environment
cp .env.example .env
# Edit .env with your credentials

# Run locally
uvicorn main:app --reload

3. Deploy to AWS EC2

# SSH into your EC2 instance
ssh -i your-key.pem ubuntu@your-ec2-ip

# Clone and deploy
git clone https://github.com/farizanjum/reply.git app
cd app

# Configure environment
cp backend/.env.example backend/.env.prod
# Edit .env.prod with production credentials

# Deploy with Docker Compose
sudo docker-compose -f docker-compose.prod.yml up -d

4. Frontend Setup

cd frontend
npm install
npm run dev

Deploy to Vercel by connecting your GitHub repository.

Features

Core Features

• Google OAuth authentication
• YouTube API integration
• Video synchronization
• Keyword-based filtering
• Auto-reply engine
• Duplicate prevention (< 1ms lookup)
• Human-like delays
• Text variation
• Quota management
• Analytics dashboard
• Real-time system health monitoring
• Notification preferences (email alerts)

Anti-Ban Strategy

1. Random Delays

   • Before reply: 0.8-3.5 seconds
   • After reply: 1.0-2.5 seconds
   • Between batches: 2-5 minutes

2. Batch Processing

   • Random batch size: 8-15 comments
   • Randomized intervals

3. Reply Variation

   • Random greetings
   • Random closings
   • Template selection

4. Gradual Ramp-up

   • Week 1: 10 replies/day
   • Week 2: 25 replies/day
   • Week 3: 50 replies/day
   • Week 4: 100 replies/day
   • Week 5+: 180 replies/day

Configuration

Video Settings

Each video can be configured with:

{
  "auto_reply_enabled": true,
  "keywords": ["MCP", "help", "tutorial"],
  "reply_templates": [
    "Hey {name}! Check out this resource: {link}",
    "Thanks for watching! Here's what you need: {link}"
  ],
  "schedule_type": "hourly"
}

Environment Variables

Backend:

• DATABASE_URL - PostgreSQL connection (AWS RDS)
• REDIS_URL - Redis connection
• GOOGLE_CLIENT_ID - OAuth client ID
• GOOGLE_CLIENT_SECRET - OAuth secret
• YOUTUBE_API_KEY - YouTube API key
• SECRET_KEY - JWT secret
• FRONTEND_URL - Frontend URL (https://tryreply.app)
• REDIRECT_URI - OAuth redirect

Frontend:

• DATABASE_URL - Prisma database URL
• NEXT_PUBLIC_BACKEND_URL - Backend API URL
• BETTER_AUTH_SECRET - Auth secret
• GOOGLE_CLIENT_ID / GOOGLE_CLIENT_SECRET - OAuth credentials

Performance Metrics

• Duplicate Check: < 1ms per lookup
• Batch Check: 2-5ms for 100 comments
• Reply Speed: 100 comments in ~3 minutes
• Daily Capacity: 150-180 replies/day (within quota)
• Uptime: 99.9% (AWS SLA)

Project Structure

reply-comments/
├── backend/                 # FastAPI backend
│   ├── main.py             # App entry point
│   ├── config.py           # Configuration
│   ├── database_pg.py      # PostgreSQL functions
│   ├── routers/            # API routes
│   ├── services/           # Business logic
│   ├── middleware/         # Auth middleware
│   └── docker-compose.prod.yml
├── frontend/               # Next.js frontend
│   ├── src/app/           # App Router pages
│   ├── src/components/    # React components
│   └── src/lib/           # Utilities
└── README.md               # This file

Security

• OAuth 2.0 authentication
• JWT token-based sessions
• Environment variable secrets
• HTTPS only (Caddy reverse proxy)
• CORS protection
• SQL injection prevention (parameterized queries)

Monitoring

View Logs

# SSH into EC2
ssh -i your-key.pem ubuntu@your-ec2-ip

# View all container logs
sudo docker-compose -f docker-compose.prod.yml logs --tail=100

# View specific service
sudo docker-compose -f docker-compose.prod.yml logs web --tail=50

Health Check

curl https://your-api-url/health
# Returns: {"status":"healthy","postgres":true,"redis":true}

Database

# Connect to RDS PostgreSQL
psql -h your-rds-endpoint -U postgres -d youtube_autoreply

Troubleshooting

Common Issues

1. OAuth Error

   • Verify redirect URI matches Google Console
   • Check client ID and secret

2. Database Connection Error

   • Verify DATABASE_URL in .env.prod
   • Check RDS security group allows EC2 access

3. Quota Exhausted

   • Check daily usage in analytics
   • Wait for midnight PT reset

4. Container Not Starting

   • Check logs: docker-compose logs web
   • Verify environment variables

v3.0 Roadmap

Planned Features

1. AI Video Analysis

• Automatically analyze video content
• Smart link recommendations based on context
• Auto-insert relevant links in replies

2. Instagram DM Auto-Reply

• Automated responses to Instagram DMs
• Keyword-based filtering
• Template-based personalized replies
• Analytics and engagement tracking

3. One-Click Instagram Connect

• Seamless Instagram Business account connection
• Facebook OAuth integration
• Real-time DM webhook notifications

4. Google OAuth Verification

• Complete app verification for production
• Remove "unverified app" warning

Timeline (Estimated)

Phase	Features	Duration
Phase 1	AI Video Analysis	3 weeks
Phase 2	Instagram Connect	2 weeks
Phase 3	Instagram DM Auto-Reply	2 weeks
Phase 4	Testing & Polish	1 week

License

MIT License - see LICENSE file

Contributing

1. Fork the repository
2. Create feature branch (git checkout -b feature/amazing-feature)
3. Commit changes (git commit -m 'Add amazing feature')
4. Push to branch (git push origin feature/amazing-feature)
5. Open Pull Request

---

Built with ❤️ for YouTube creators