Skip to content

CHATBOT_README.md

Latest commit

 

History

History
71 lines (55 loc) · 1.87 KB

CHATBOT_README.md

File metadata and controls

71 lines (55 loc) · 1.87 KB

🤖 Chatbot Response Engine

How It Works

The chatbot uses OpenAI GPT-3.5-turbo for intelligent responses to user queries. When OpenAI API is unavailable, it falls back to predefined responses for common questions.

Response Flow:

  1. User sends a message
  2. System categorizes the message (greeting, pricing, support, features)
  3. Analyzes sentiment (positive, negative, neutral)
  4. Calls OpenAI API if available, otherwise uses predefined responses
  5. Saves conversation to database
  6. Returns response to user

Message Categories:

  • greeting: hello, hi, hey
  • pricing: price, cost, pricing
  • support: help, support, issue, problem
  • features: feature, capability, what can
  • default: other queries

Predefined Responses:

When OpenAI is unavailable, the chatbot uses simple predefined responses:

greeting: "Hello! How can I help you today?"
pricing: "Our pricing plans start at $9.99/month..."
support: "I'm here to help! Can you describe the issue?"
features: "Our platform offers mass email sending..."

Setup

  1. Add OpenAI API key to backend/.env:
OPENAI_API_KEY=your_openai_api_key_here
  1. Install dependencies:
cd backend && npm install
cd ../frontend && npm install
  1. Start servers:
# Backend
cd backend && npm run dev

# Frontend  
cd frontend && npm run dev
  1. Visit /chatbot or use the chat widget

API Endpoints

  • POST /api/chatbot/message - Send message, get response
  • GET /api/chatbot/history/:userId - Get chat history
  • POST /api/chatbot/test - Test chatbot

Files Created

Backend:

  • services/chatbotService.js - Core AI logic
  • controllers/chatbotController.js - API handlers
  • routes/chatbotRoutes.js - Routes
  • models/Message.js - Database schema

Frontend:

  • pages/Chatbot.jsx - Chat interface
  • components/ChatbotWidget.jsx - Embeddable widget