Context Bot (RAG Application)

A Next.js app for Retrieval-Augmented Generation (RAG) using OpenAI, Qdrant, and LangChain. Index and query data from PDFs, websites, and text.

Features

• Upload and index PDFs, CSVs, and text files
• Index website content and remote PDFs
• Store embeddings in Qdrant (via Docker)
• Chat interface powered by OpenAI (GPT-4o-mini, GPT-4.1-mini)
• Context-aware answers based on your uploaded sources
• Modern UI with Tailwind, Radix, and Lucide icons
• Theme support (light/dark)

Demo

Watch the demo on YouTube: https://youtu.be/Zi-qxFximXI

Screenshots

Project Structure

app/
  api/                # REST endpoints for chat, file/text/website storage, source listing/deletion
  components/         # UI components (chat window, file upload, etc.)
  lib/                # Utility functions, text splitting, store logic
  public/             # Static assets (SVGs, favicon, screenshots)
  uploads/            # Uploaded files (created at runtime)
scripts/              # Utility scripts (e.g., test-qdrant.cjs)

Getting Started

1. Start Qdrant (vector DB) via Docker:

   docker-compose up -d

2. Install dependencies:

   npm install

3. Run the development server:

   npm run dev

4. Open http://localhost:3000

API Endpoints

• POST /api/chat — Ask questions, get context-aware answers
• POST /api/store-file — Upload and index files (PDF, CSV, TXT)
• POST /api/store-text — Store raw text
• POST /api/store-website — Index website or remote PDF
• GET /api/list-sources — List indexed sources
• POST /api/delete-source — Delete a source by ID

Tech Stack

• Next.js, React, TypeScript
• LangChain, OpenAI, Qdrant
• Tailwind CSS, Radix UI, Lucide Icons
• Docker (Qdrant)

---