Chatbot

• Documents-based AI Assistant to help users
• RAG (Retrieval-Augmented Generation) architecture for accurate technical responses
• Hybrid Approach (Local Processing + Cloud-based Vector DB and LLM Service)

Tech Stack

• AI: Groq SDK - Llama 3.1 8B Instant Model (Cloud-based)
• Database: Pinecone Vector Database (Cloud-based)
• Embedings: Transformer.js with Xenova/all-MiniLM-L6-v2 Model (Local)
• Frontend: React, TypeScript, Tailwind CSS
• Backend: Electron, Node.js
• Build: Vite, Electron Builder

Quick Start

Prerequisites

• Node.js (v16 or higher)
• npm or yarn
• API keys for services (see Configuration below)

Steps

• Clone the repository

  git clone https://github.com/swetha-nbase2/chatbot.git

• Install dependencies

  npm install

• Setup the API Keys

  Configure your .env file with API keys

  cp .env.example .env

  Edit .env and add your API keys:

  # Groq API Configuration
  GROQ_API_KEY = your_groq_api_key_here
  
  # Pinecone Configuration
  PINECONE_API_KEY = your_pinecone_api_key_here
  PINECONE_ENVIRONMENT = your_pinecone_environment_here
  PINECONE_INDEX_NAME = your_pinecone_index_name_here
  
  # Chat Configuration
  MAX_CHAT_HISTORY = max_chat_history_here
  EMBEDDING_DIMENSION = embedding_dimension_here

  Create Groq API Key:

  1. Visit Groq Console
  2. Create an account and generate an API key
  3. Add to .env as GROQ_API_KEY

  Create Pinecone API Key:

  1. Visit Pinecone
  2. Create an account and get your API key
  3. Note your environment region
  4. Add to .env as PINECONE_API_KEY and PINECONE_ENVIRONMENT

  Chat Configuration:

  1. Recommended 20 Maximum Chat History
  2. Embedding Dimension of Xenova/all-MiniLM-L6-v2 is 384

• Process documents (one time)

  Create a folder called 'documents' inside the root directory and place the documents/data sheets (supported file formats - .txt, .pdf, .md) that are going to serve as knowledge base for the chatbot.

  npm run process-docs

• Start the app

  npm run electron-dev

The app uses an optimized workflow:

1. One-time document processing: Run npm run process-docs once to prepare your knowledge base
2. Fast app startup: Use npm run electron-dev for instant startup
3. Update when needed: Re-run document processing only when you add new documents

Usage

1. Start the application
2. Click the chatbot button (🤖) in the bottom-right corner
3. Type your message and press Enter or click Send
4. Enjoy AI-powered conversations with context memory!

Workflow

Knowledge Preparation:

• Upload necessary documents and data sheets
• Documents are chunked into smaller text segments
• Each chunk is embedded
• Embeddings are stored in vector database
• Knowledge base is ready for user queries

Query Processing:

• User submits a query through chat interface
• Query is embedded into vector
• Vector DB retrieves the most relevant context chunks
• Context chunks and user query are combined into structured prompt
• LLM generates natural language response
• Response is returned to user in chat interface

Important Scripts

• Process the Documents

  npm run process-docs

• Clear the Database (Must before processing the new documents)

  npm run clear-db

• Transpile (Must after making changes in electron files)

  npm run transpile

• Start the React Server

  npm run dev:react

• Start the App

  npm run electron-dev

License

This project is licensed under the MIT License.

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test thoroughly
5. Submit a pull request

---

Happy Chatting!