RAG Chatbot with Multi-Tool Integration

Overview

An intelligent AI-powered chatbot built with LangChain, Google Gemini API, and RAG (Retrieval-Augmented Generation). This project integrates multiple tools including document search, SQL database queries, and web scraping.

Status: AAIDC Module 1 Submission

Features

✨ Core Capabilities:

• Retrieval-Augmented Generation (RAG): Search through personal documents using semantic search
• Multi-Tool Integration: Document search, SQL database queries, web scraping
• Smart Intent Recognition: Automatic detection of user intent to use appropriate tools
• Conversation Memory: Maintains conversation context across multiple turns
• Web Interface: Clean, interactive chat UI built with Flask

System Architecture

User Query → Intent Analysis → Relevant Tools Activated → Context Building → LLM Response Generation → Response to User

Installation

Prerequisites

• Python 3.8 or higher
• Google Gemini API key (free from Google AI Studio)
• 4GB RAM minimum

Setup

1. Clone the repository:

git clone https://github.com/Manish-010/rag-chatbot-aaidc.git
cd rag-chatbot-aaidc

2. Create virtual environment:

python -m venv venv
venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Setup environment variables:

cp .env.example .env
# Edit .env with your Gemini API key

5. Run the application:

python app/run_server.py

Access at: http://localhost:5000

Usage

Example Queries

Document Search:

• "What is the helmet detection project about?"
• "Tell me about the RAG agent report"

Database Queries:

• "How many users are in the sales department?"
• "Show me all employees in engineering"

Web Queries:

• "What's on https://example.com?"

Project Structure

rag-chatbot-aaidc/
├── src/
│   ├── chatbot_logic.py
│   └── config.py
├── app/
│   ├── app.py
│   └── run_server.py
├── static/
├── templates/
├── data/sample_documents/
├── tests/
├── requirements.txt
├── .env.example
├── README.md
└── LICENSE

How It Works

1. Document Processing (RAG): Loads .docx files, creates embeddings, stores in FAISS
2. Intent Recognition: Analyzes user query for keywords
3. Response Generation: Builds context and sends to Gemini API

Evaluation Criteria Met

✅ Functionality: Complete RAG pipeline with multi-tool integration
✅ Reproducibility: Clear setup instructions
✅ Documentation: Architecture and usage explained
✅ Code Quality: Modular design with error handling
✅ Demo: Web interface with working examples

GitHub Repository

Full code and documentation: https://github.com/Manish-010/rag-chatbot-aaidc

License

MIT License