ReplyTrap

Free and open source AI-powered scambaiting suite that monitors email accounts, detects scams, and uses customizable personas to engage scammers in conversations that waste their time.

Demo

replytrap-demo.mp4

Quick Start

Docker Setup

Option 1: Docker Desktop GUI - recommended for beginners

Install Docker Desktop for Windows/Mac: Download Docker Desktop

1. Open Docker Desktop
2. Navigate to the reply_trap folder in your file explorer
3. Right-click on docker-compose.yml → "Compose Up"

Option 2: Terminal Install Docker your preferred way

git clone https://github.com/yourusername/reply_trap.git
cd reply_trap
docker-compose up

⚠️ Docker notes:

• First run takes several minutes to download the llama3.2:3b model (around 3 GB)
• Can be resource-intensive when LLMs are running since the AI model runs locally on your computer - this eliminates external API costs and dependencies
• Runs in single user mode by default (no login required) for easy setup. Change SINGLE_USER_MODE to false in docker-compose.yml to enable user registration and login

Local Setup

git clone https://github.com/yourusername/reply_trap.git
cd reply_trap
cp .env.example .env
# Edit .env - choose your LLM provider and delete the others
bin/setup
# In another terminal: bin/jobs

Visit http://localhost:3030

Key Features

• AI persona generation - write your own personas or generate them with AI
• Automated email monitoring and scam detection
• Smart scam responses - answers stay consistent with your persona and timing varies from minutes to 2 days to mimic human behavior
• Auto-testing email credentials - makes setting up your email accounts straightforward
• Encrypted email credentials storage
• Ollama integration for local AI models - Docker only. Saves configuration effort and cost
• Disable email accounts or conversations you don't want to continue

Disclaimer

Educational and defensive security use only. Use responsibly and legally.