RAG-PDF Chatbot 🤖📄

A document-based AI assistant that allows users to upload a PDF file and chat with an AI bot capable of answering questions based on the uploaded content. Built with modern web technologies and a robust cloud-native backend.

---

🔍 Description

This app mimics a basic AI PDF Q&A assistant in two steps:

1. Screen 1: User enters their email to simulate login (stored in localStorage).
2. Screen 2: A chat UI appears, but interaction is disabled until a PDF is uploaded.

Key functionality includes:

• Upload a single PDF file (max 10MB).
• File is sent to AWS S3 via a presigned URL.
• An AWS Step Functions workflow is triggered automatically:
  • Extracts text
  • Chunks it
  • Embeds chunks
  • Indexes embeddings in Pinecone
• Once processing succeeds, user can chat with an AI based on the document contents.

---

📁 Project Structure

.
├── client/             # React + Vite frontend
├── server/             # Express backend API
└── lambda-functions/   # AWS Lambda logic used by Step Functions

---

🚀 Features

• 🔐 Email-based (localStorage) access
• 📄 Single-file upload (PDF, max 10MB)
• 🧠 AI Q&A based on uploaded document
• ☁️ AWS S3 + Step Functions integration
• 📚 Pinecone vector search
• ✨ Modern and sleek chat UI

---

🛠️ Tech Stack

Frontend

• React 19 + Vite
• TailwindCSS 4
• Shadcn UI + Lucide Icons
• TanStack Query (polling for file processing)
• React Router

Backend

• Express 5
• AWS SDK (S3, DynamoDB, Step Functions)
• Pinecone Vector DB
• Gemini API

---

⚙️ Installation

1. Clone the Repository

git clone https://github.com/andrew-dev-p/AWS-RAG-PDF-chat
cd aws-rag-pdf-chat

2. Setup Client

cd client
npm install

3. Setup Server

cd ../server
npm install

---

🧪 Running Locally

Start Client

cd client
npm run dev

Start Server

cd server
npm run dev

---

🔐 Environment Variables

Client (client/.env.local)

VITE_API_URL=http://localhost:3000/api

Server (server/.env)

PORT=3000
NODE_ENV=development

AWS_REGION=us-east-1
AWS_BUCKET_NAME=aws-rag-pdf-chat
AWS_ACCESS_KEY_ID=your_access_key_id
AWS_SECRET_ACCESS_KEY=your_secret_access_key

PINECONE_API_KEY=your_pinecone_api_key
GEMINI_API_KEY=your_gemini_api_key

---

⚙️ Deployment

• Frontend: Hosted on Vercel
• Backend: Deployed on Render
• Storage: Files saved to AWS S3
• Processing: AWS Step Functions + Lambda
• Vector DB: Pinecone
• AI Responses: via Gemini API