Project Overview

This repository contains a full-stack RAG chatbot powered by LangChain and IONOS, with separate frontend and backend folders:

• frontend: A Next.js (React) application that allows users to input a page URL, select a model, and chat with an AI assistant based on website content.

• backend: A FastAPI service that:

  1. Scrapes and indexes webpage text using TF-IDF for RAG.
  2. Routes chat requests to IONOS AI models, managing conversation history.
  3. Exposes endpoints for initializing RAG index, fetching chat history, and sending user messages.

---

Table of Contents

• Prerequisites
• Environment Variables
• Backend Setup
• Frontend Setup
• Usage
• Project Structure
• License

---

Prerequisites

Before you begin, ensure you have:

• Node.js (v18 or above) and npm or yarn
• Python (v3.10 or above)
• pip or poetry for Python dependencies
• An IONOS API Key for language model access

---

Environment Variables

Create a .env file in both the frontend and backend folders using the following template: (you can instead create a single unique .env file at project root if you prefer)

# Frontend (Next.js)
NEXT_PUBLIC_APP_BASE_URL=http://localhost:8000  # URL of the backend API

# Shared / Backend (.env)
IONOS_API_KEY=your_ionos_api_key_here           # IONOS AI Model Hub key
RAG_K=3                                         # top-k RAG chunks to retrieve (default: 3)
CHUNK_SIZE=500                                  # chars per chunk (default: 500)
MAX_CHUNK_COUNT=256                             # maximum number of chunks (default: 256)

• NEXT_PUBLIC_APP_BASE_URL: URL where your backend is running, used by the frontend.

• IONOS_API_KEY: Your secret key for accessing IONOS AI Model Hub (required by the backend).

• RAG_K: Number of top chunks to retrieve for context.

• CHUNK_SIZE: Maximum characters per chunk when splitting scraped text.

• MAX_CHUNK_COUNT: Cap on total chunks to index.

---

Backend Setup

1. Navigate to the backend folder:

   cd backend

2. Install dependencies:

   pip install -r requirements.txt

3. Run the FastAPI server locally:

   python main.py

   The backend will be available at http://localhost:8000.

---

Frontend Setup

1. Navigate to the frontend folder:

   cd frontend

2. Install dependencies:

   npm install
   # or
   yarn install

3. Start the development server:

   npm run dev
   # or
   yarn dev

   The frontend will be available at http://localhost:3000.

---

Usage

1. Open your browser at http://localhost:3000.
2. Enter a page URL to scrape and wait for RAG initialization.
3. Select an AI model from the dropdown.
4. Start chatting—messages will be sent to the backend, enriched with top-k context, and answered by your chosen model.

---

Project Structure

├── frontend             # Next.js React app
│   ├── .env             # Frontend environment config
│   ├── app/
│   ├── components/
│   ├── public/
│   ├── lib/
│   └── package.json
└── backend              # FastAPI service
    ├── .env             # Backend environment config
    ├── main.py          # FastAPI entrypoint
    ├── requirements.txt
    └── other modules…

---

License

This project is released under the MIT License. Feel free to use and modify it in your own applications.