RAG Chatbot (Website / PDF / Text)

A Streamlit-based Retrieval-Augmented Generation (RAG) chatbot that answers questions from websites, PDFs, or raw text using embeddings, FAISS vector search, and a Groq-powered LLM.

---

Features

• Website-based Q&A
• PDF-based Q&A
• Text-based Q&A
• Chat-style UI with history
• New Chat (session reset)
• Apply button (runs only on click)
• No hallucinations (context-only answers)
• Deployment-ready (Streamlit Cloud / AWS)

---

How It Works (RAG Flow)

1. User selects Website / PDF / Text
2. Content is ingested and split into chunks
3. Chunks are embedded and stored in FAISS
4. User asks a question
5. Relevant chunks are retrieved
6. LLM generates an answer strictly from context
7. Answer is shown in Streamlit UI

---

Project Structure

│
├── app.py # Streamlit UI (Frontend)
├── main.py # Backend Orchestrator
├── requirements.txt
├── README.md
├── .env
├── .gitignore
│
├── embeddings/ # FAISS vector database (auto-created)
├── logs/ # Application logs
├── uploaded_pdfs/ # Uploaded PDF files (optional use)
│
├── src/
│ ├── pycache/
│ │
│ ├── components/
│ │ ├── init.py
│ │ └── ragchatbot.py # RAG + Groq LLM logic
│ │
│ ├── datatransformer/
│ │ ├── init.py
│ │ ├── webdatatransfer.py # Website text extraction
│ │ ├── textdatatransfer.py# Text & PDF text splitting
│ │ └── pdfdatatransfer.py # (Optional PDF logic)
│ │
│ └── utils/
│ ├── init.py
│ ├── dataembedding.py # Embedding creation
│ └── dataingestion.py # Website ingestion logic

Installation & Setup

Clone the Repository

git clone https://github.com/kumar-kiran-24/chatbot
pip install -r requirments.txt

cd chatbot

streamlit run app.py

env file

.env

GROQ_API_KEY=your_groq_api_key_here

for test

https://chatbot-cah6sgfpqmndtmydp7hya4.streamlit.app/