📄 AI Document Q&A Chatbot (RAG Pipeline)

A full-stack AI-powered chatbot that lets you upload any PDF and ask questions about it in plain English. Built with a Retrieval-Augmented Generation (RAG) pipeline using LangChain, ChromaDB, and Anthropic's Claude LLM — deployed on AWS.

🧠 How It Works

Instead of asking an LLM to guess answers, this app first retrieves the most relevant content from your document, then passes it to Claude as context — so every answer is grounded in what's actually in the PDF.

Upload PDF → Extract & Chunk Text → Generate Embeddings → Store in ChromaDB
                                                                    ↓
                                          User Question → Semantic Search → Claude → Answer

---

🛠️ Tech Stack

Layer	Technology
Language	Python 3.10+
Backend API	FastAPI + Uvicorn
AI Orchestration	LangChain
LLM	Anthropic Claude API
Vector Database	ChromaDB
Embeddings	HuggingFace (all-MiniLM-L6-v2)
File Storage	AWS S3
Deployment	AWS EC2 (Ubuntu)
Frontend	Streamlit
Environment	python-dotenv

---

📁 Project Structure

rag-chatbot/
├── app/
│   ├── __init__.py
│   ├── ingest.py       # PDF loading, chunking, ChromaDB storage
│   ├── rag.py          # Semantic search + Claude integration
│   └── main.py         # FastAPI endpoints
├── uploads/            # Temporary local storage before S3 upload
├── .env                # API keys (never committed)
├── .gitignore
├── requirements.txt
└── streamlit_app.py    # Frontend chat UI

---

⚙️ Getting Started

Prerequisites

• Python 3.10+
• An Anthropic API key
• An AWS account with an S3 bucket

1. Clone the repo

git clone https://github.com/xMayble/rag-chatbot.git
cd rag-chatbot

2. Create a virtual environment

python -m venv venv

# Windows
venv\Scripts\activate

# Mac/Linux
source venv/bin/activate

3. Install dependencies

pip install -r requirements.txt
pip install sentence-transformers

4. Set up environment variables

Create a .env file in the root directory:

ANTHROPIC_API_KEY=your_anthropic_api_key
AWS_ACCESS_KEY_ID=your_aws_access_key
AWS_SECRET_ACCESS_KEY=your_aws_secret_key
AWS_BUCKET_NAME=your_s3_bucket_name
AWS_REGION=us-east-1

5. Run the app

Terminal 1 — Start the API:

uvicorn app.main:app --reload

Terminal 2 — Start the frontend:

streamlit run streamlit_app.py

Then open http://localhost:8501 in your browser.

---

📡 API Endpoints

Method	Endpoint	Description
GET	/	Health check
POST	/upload	Upload and ingest a PDF
POST	/ask	Ask a question about the document

---

☁️ AWS Infrastructure

• S3 — Stores uploaded PDFs for scalable, production-grade file handling
• EC2 (t2.micro, Ubuntu) — Hosts the FastAPI backend with a live public URL

---

📌 Key Concepts

RAG (Retrieval-Augmented Generation) — A technique that improves LLM accuracy by retrieving relevant context from a knowledge base before generating a response. This prevents hallucinations and grounds answers in real document content.

Vector Embeddings — Text chunks are converted into numerical vectors that capture semantic meaning. ChromaDB stores these and enables similarity search — finding relevant chunks by meaning, not just keywords.

---

🔮 Future Improvements

• Multi-PDF support
• Chat history / memory
• Source citations showing which page the answer came from
• User authentication
• Frontend rebuild in React

---

👤 Author

Mehbub Rohit
LinkedIn · GitHub