Live at : http://217.154.38.177:8501/

Smart ChatBot 🤖

This is a simple Streamlit ChatBot app powered by OpenAI, LangChain, RAG (Retrieval-Augmented Generation), and FAISS for vector storage.
The app allows users to upload a text-based PDF document and ask natural language questions related to the content of the uploaded file.

🚀 Features

🔥 Ask questions from any uploaded PDF document
💡 Uses OpenAI (gpt-3.5-turbo) to generate smart answers
🔎 Powered by LangChain’s RAG framework and FAISS vector store
📚 Dynamically parses PDFs and generates embeddings
🧠 Retrieval-based contextual answering with document chunking

🛠️ Tech Stack

Streamlit – UI frontend
LangChain – Chain and retrieval logic
OpenAI – LLM for answer generation
FAISS – Local in-memory vector storage
PDF document parsing via PyPDFLoader

📦 Folder Structure

your-repo/
├── app.py             # Main Streamlit app file
└── README.md           # Project documentation

💡 How It Works

Upload a PDF document (text-based only).
It gets split into chunks using LangChain’s RecursiveCharacterTextSplitter.
Each chunk is converted into a vector using OpenAI Embeddings (text-embedding-3-small or similar).
Vectors are stored in a temporary FAISS index.
At query time, most relevant chunks are retrieved and passed as context to GPT.
GPT returns an answer grounded in the uploaded document.

▶️ Getting Started

1. Clone the repository

git clone https://github.com/siddharthsingh5010/pdf_rag_chatbot
cd pdf_rag_chatbot

2. Install dependencies

Create a virtual environment and install required packages:

pip install -r requirements.txt

3. Set your OpenAI API Key

You can export it in your terminal session:

docker build -t pdf_rag_app .

Or add this in your shell config (~/.bashrc, ~/.zshrc, etc.).

4. Run the app

docker run -p 8501:8501 pdf_rag_app

🔐 Notes

This app only supports text-based PDFs (not scanned images).
For best performance, make sure your OpenAI API key has access to gpt-3.5-turbo.

📄 License

MIT License

Author
Siddharth Singh

PDF RAG ChatBot