merendamattia/amazon-ai-agent repository overview

⁠📝 Amazon AI Agent

An intelligent AI agent that generates detailed and professional Amazon product reviews and sales listings from product links. Powered by OpenAI and Google AI models with advanced web scraping and content analysis capabilities.

⁠✨ Features

• 🤖 Multi-Provider Support - Works with OpenAI (GPT) and Google (Gemini) models
• 📝 Product Reviews - Generate detailed and professional Amazon product reviews
• 💼 Sales Listings - Create persuasive sales listings for platforms like eBay, Subito, or Facebook Marketplace
• 🔗 Direct Link Processing - Paste an Amazon product link and get instant output
• 📱 Telegram Bot - Interact with the agent directly through Telegram
• 🌐 Web Content Extraction - Advanced web fetching with intelligent token management
• 📊 Smart Token Management - Automatic content truncation for optimal performance
• 💬 Multi-Tool Agent System - Uses DataPizza framework for powerful agent capabilities
• 🛡️ Error Handling - Robust error handling and logging throughout
• ⚙️ Flexible Configuration - Easy switching between providers and models
• 🔄 URL Expansion - Automatic expansion of shortened Amazon URLs (amzn.to, a.co, etc.)

⁠📋 Requirements

• Python 3.11+
• AI Provider API Key (OpenAI or Google)
• Telegram Bot Token (optional, only for Telegram bot mode)

⁠🚀 Getting Started

⁠Installation

1. Clone the repository and enter the directory:

   git clone https://github.com/merendamattia/amazon-ai-agent.git
   cd amazon-ai-agent
   

   Copy

2. Create and activate a Python virtual environment:

   conda create --name amazon-ai-agent-env python=3.11.13
   conda activate amazon-ai-agent-env
   

   Copy

   Or using venv:

   python3 -m venv venv
   source venv/bin/activate
   

   Copy

3. Install dependencies:

   python -m pip install --upgrade pip
   pip install -r requirements.txt
   

   Copy

4. Configure environment variables:

   cp .env.example .env
   # Edit .env with your preferred AI provider API key
   

   Copy

   For Google (default provider):

   GOOGLE_API_KEY=your_api_key_here
   GOOGLE_MODEL=gemini-2.5-flash
   TELEGRAM_BOT_TOKEN=your_telegram_token_here  # Optional
   

   Copy

   For OpenAI:

   OPENAI_API_KEY=your_api_key_here
   OPENAI_MODEL=gpt-4o-mini
   TELEGRAM_BOT_TOKEN=your_telegram_token_here  # Optional
   

   Copy

⁠Running the Application

⁠CLI - Generate a review or sales listing for an Amazon product:

# Generate a review using default provider (Google)
python app.py https://www.amazon.com/your-product-link

# Generate a sales listing
python app.py --type listing https://www.amazon.com/your-product-link

# Use OpenAI provider explicitly
python app.py --provider openai https://www.amazon.com/your-product-link

# Generate listing with OpenAI
python app.py --type listing --provider openai https://www.amazon.com/your-product-link

Supported providers:

• google (default) - Uses Google's Gemini models
• openai - Uses OpenAI's GPT models

⁠Telegram Bot - Interactive chat interface:

Quick start:

# 1. Update .env with TELEGRAM_BOT_TOKEN and your preferred provider credentials
# 2. Run the bot with default provider (Google)
python telegram_bot.py

# Or specify a different provider
python telegram_bot.py --provider openai

Then open Telegram, find your bot, and start using it!

Bot Features:

• 📝 Interactive menu to generate reviews or sales listings
• 🔗 Simply paste Amazon links
• ⏳ Real-time processing status updates
• 💬 Support for long outputs (automatically split into multiple messages)
• ❌ Easy cancel option
• 💼 Choose between product reviews and sales listings

⁠Docker

Telegram Bot Mode:

docker build -t amazon-ai-agent:local .
docker run --env-file .env --restart unless-stopped amazon-ai-agent:local

Docker Compose (easiest):

# Run Telegram Bot
docker compose up

⁠🤝 Contributing

We welcome contributions! Please see CONTRIBUTING.md⁠ for:

• Development setup
• Code style guidelines
• Commit conventions
• Pull request process

⁠Troubleshooting

⁠Common Issues

"API key not found in .env"

• Ensure your .env file exists in the project root directory
• Verify you have configured the credentials for your chosen provider
• Check that the environment variable names are correct

"Unsupported provider" error

• Ensure you're using a valid provider: openai or google
• Verify that the corresponding API credentials are set in your .env file

"Connection timeout"

• Verify your internet connection
• Check the service status for your provider (OpenAI or Google)

"Rate limit exceeded"

• Wait a moment before retrying
• Consider using a more cost-effective model variant

⁠📄 License

This project is licensed under the Apache License 2.0 - see the LICENSE⁠ file for details.

⁠📞 Support

For issues, feature requests, or questions:

• Open an issue on GitHub⁠

⁠🙏 Acknowledgments

• Built with DataPizza AI Framework⁠
• Powered by OpenAI⁠