POPBOT

A modern, AI chatbot powered by Google's Gemini model with GraphRAG knowledge retrieval, built with Next.js, Neo4j, Clerk, Convex, and shadcn/ui.

Features

• 🤖 Powered by Google's Gemini AI model
• 🧠 GraphRAG-powered knowledge retrieval using Neo4j
• 🔐 User authentication with Clerk
• 💾 Real-time data sync with Convex
• 🎨 Beautiful UI with shadcn/ui components
• 📱 Responsive design
• 🔌 Easy to embed in any website
• ⚡ Real-time chat interactions

Tech Stack

• Framework: Next.js 15
• AI: Google Gemini, LangChain
• Knowledge Graph: Neo4j (GraphRAG)
• Auth: Clerk
• Database: Convex
• UI: shadcn/ui, Tailwind CSS
• Language: TypeScript

Getting Started

1. Clone the repository

2. Install dependencies:

   npm install

3. Create a .env.local file in the root directory and add your Gemini API key:

   NEXT_PUBLIC_GEMINI_API_KEY=your_api_key_here
   

4. Run the development server:

   npm run dev

5. Open http://localhost:3000 to see the demo.

Embedding the Chatbot

To embed the chatbot in your website, add the following iframe to your HTML:

<iframe 
  src="https://embeddable-bot.vercel.app/widget"
  width="100%"
  height="600px"
  frameBorder="0"
/>

Using the Chat Widget

The chat widget can be imported and used in any React component:

import { ChatWidget } from '@/components/ChatWidget';

export default function YourPage() {
  return (
    <div>
      <h1>Your Page Content</h1>
      <ChatWidget />
    </div>
  );
}

Document Limits

The app processes one document at a time. For best results:

Document Size	Recommendation
10-20 pages	Optimal accuracy
30-50 pages	Acceptable
50+ pages	Split into sections

Technical constraints

• Graph build timeout: 60 seconds max processing time
• Chunk size: ~1500 characters per chunk
• Context retrieval: 8 chunks + 30 facts per query
• Supported formats: PDF, TXT, Markdown

For larger documents, split them into focused sections and upload one at a time.

Development

• src/components/Chatbot.tsx - Main chatbot component
• src/components/ChatWidget.tsx - Floating chat widget
• src/lib/gemini.ts - Gemini AI configuration
• src/lib/graph-rag.ts - GraphRAG knowledge extraction
• src/app/widget/page.tsx - Embeddable widget page

License

MIT