AI Chatbot Platform (RAG + Normal Chatbot)

A full-stack AI chatbot system that supports both:

• Normal Chatbot (LLM-based)
• RAG Chatbot (Website / PDF / Text-based Q&A)

Built using FastAPI + Next.js + FAISS + Sentence Transformers + Groq LLM, and deployed on Hugging Face Spaces (Backend) and Vercel (Frontend).

---

🌐 Live Demo

• Frontend (Vercel):
  https://chatbot-frontend-bice-nu.vercel.app/rag-chatbot

• Backend API (Hugging Face):
  https://kirankumar29-chatbot.hf.space

---

Features

Normal Chatbot

• Conversational AI (Groq LLM)
• Session-based chat
• ChatGPT-like UI

RAG Chatbot

• Website-based Q&A
• PDF-based Q&A
• Text-based Q&A
• Context-aware answers (no hallucination)
• FAISS vector search

Frontend UI

• Modern ChatGPT-style dark theme
• Sidebar navigation
• Separate Normal & RAG chatbot pages
• Upload + Chat workflow
• Real-time responses

---

How It Works (RAG Flow)

1. User uploads data (Website / PDF / Text)
2. Content is split into chunks
3. Embeddings are generated
4. Stored in FAISS vector database
5. User asks a question
6. Relevant chunks are retrieved
7. Groq LLM generates answer using context

---

Tech Stack

🔹 Backend

• FastAPI
• LangChain
• FAISS
• Sentence Transformers
• Groq API
• Qdrant

🔹 Frontend

• Next.js (React)
• Tailwind CSS
• shadcn/ui
• Axios

🔹 Deployment

• Hugging Face Spaces (Backend)
• Vercel (Frontend)

---

Project Structure

.
├── backend/
│   ├── app.py
│   ├── main.py
│   ├── requirements.txt
│   ├── uploads/
│   ├── embeddings/
│   └── src/
│       ├── components/
│       │   ├── chatbot.py
│       │   └── ragchatbot.py
│       ├── datatransformer/
│       └── utils/
│
│
├── Dockerfile
├── .env
└── README.md

Installation Setup & Run

🔹 Clone Repository

git clone https://github.com/kumar-kiran-24/chatbot
cd chatbot
pip install -r requirements.txt
uvicorn app:app --reload

Environment Variables

.env

QDRANT_URL="qdrant url"
QDRANT_API_KEY="your api key"
HF_TOKEN="your api key"
GROQ_API_KEY="your api key"

API Endpoints

• Endpoint Description
• /chatbot Normal chatbot
• /ragchatbot_url RAG chatbot
• /upload-pdf Upload PDF
• /web Load website
• /text Add raw text