AI Email Reply Agent

A Streamlit application that automatically processes and generates responses to customer emails using AI. The app uses LangChain and Groq for natural language processing and understanding.

Key Features

• Email Categorization: Automatically categorizes emails into specific categories like price inquiries, customer complaints, product inquiries, customer feedback, and off-topic emails.
• Automatic Research: Performs web searches for complex queries or specific product details using the Tavily API.
• Professional Response Generation: Generates tailored, professional email responses using the Groq API.
• Streamlit Interface: A user-friendly web interface to input email content, view categorization, see research, and generate responses in real-time.

Project Structure

• src/: Contains the main application code
  • chains.py: Logic for categorizing emails, conducting research, and generating responses.
  • config.py: Configuration and environment variable loading.
  • processors.py: Functions for processing emails, performing research, and drafting responses.
  • prompts.py: Templates used for AI prompt generation.
  • state.py: State management for the email processing workflow.
  • workflow.py: Defines the workflow using state graphs for processing emails.
• utils/: Utility functions for file operations and other helper tasks.
• outputs/: Stores generated response files like email drafts and research info.
• .env: Environment variables for API keys (Groq, Tavily).

Setup

1. Clone the repository

   git clone https://github.com/Bhavik-Jikadara/AI-Email-Reply-Agent.git
   cd AI-Email-Reply-Agent

2. Create a virtualenv (windows user)

   pip install virtualenv
   virtualenv venv
   
   # windows users
   source venv/Scripts/activate
   
   # mac users
   source venv/bin/activate

3. Install dependencies:

   pip install -r requirements.txt

4. Create a .env file with your API keys:

   • GROQ_API_KEY
   • TAVILY_API_KEY

   GROQ_API_KEY=your_groq_api_key
   TAVILY_API_KEY=your_tavily_api_key

5. Run the Streamlit app:

   streamlit run app.py

Usage

1. Open the Streamlit app in your browser
2. Enter the email content in the text area
3. Click "Generate Response"
4. View the generated professional response

License

MIT