A next-generation research paper discovery and knowledge management system competing with NotebookLM
DataEngineX transforms how researchers discover, analyze, and connect academic papers through AI-powered insights and interactive knowledge visualization.
Create a long-form research platform that enables deep understanding of complex research domains through:
- Search-First Discovery: Ask research questions, get relevant papers instantly
- Interactive Knowledge Bases: Build and visualize connections between ideas
- AI Paper Agents: Chat with papers like Cursor AI agents for PDFs
- Connection Mapping: Visual canvas showing how concepts and papers relate
- AI-Powered Search: Semantic search through research papers
- Instant Results: Real-time paper discovery with ArXiv integration
- Smart Filtering: Impact levels, citations, topics, and years
- Demo Mode: Works without API keys using intelligent mock data
- Interactive Network: Papers as nodes, connections as edges
- Visual Exploration: Click, drag, and explore paper relationships
- Smart Clustering: Related papers automatically grouped
- Real-time Search: Filter and highlight relevant papers
- Cursor-Style Chat: AI agent for each paper
- Contextual Q&A: Ask about methodology, results, implications
- Reference Linking: Answers link to specific sections and pages
- PDF Integration: Side-by-side paper view with chat interface
- Trend Analysis: Emerging topics and research directions
- Citation Networks: Author and institutional connections
- Knowledge Gaps: Identify unexplored research areas
- Framework: Remix (React) with TypeScript
- Styling: Tailwind CSS + shadcn/ui components
- Animations: Custom CSS animations + Framer Motion
- Icons: Lucide React (modern, consistent icons)
- Framework: FastAPI with automatic OpenAPI docs
- Search: ArXiv API integration
- AI Processing: Chunkr AI for document processing
- Database: Supabase (production) / Demo mode (development)
- RAG: Semantic search within saved papers
{
"remix": "^2.16.8",
"tailwindcss": "^3.4.17",
"lucide-react": "^0.511.0",
"@headlessui/react": "^2.0.0",
"framer-motion": "^11.0.0"
}- Node.js 20+
- npm or yarn
- (Optional) Backend API running on localhost:8000
# Clone the repository
git clone <repository-url>
cd DelphiX
# Install dependencies
npm install
# Start development server
npm run dev# For full functionality, start the backend API
# Follow backend README for setup instructions
python3 -m uvicorn main:app --host 0.0.0.0 --port 8000 --reloadDemo Mode: The frontend works standalone with intelligent mock data for hackathon demos.
- Open DataEngineX → Beautiful animated landing page
- Ask Research Question: "How do transformers work in NLP?"
- Instant Results: 10+ relevant papers with impact indicators
- Visual Appeal: Gradient backgrounds, smooth animations
- Select Papers: Click papers of interest
- Create Knowledge Base: Single button click
- Navigate to Canvas: Animated transition to network view
- Explore Connections: Interactive paper nodes with relationships
- Click Paper Node: Select "Attention Is All You Need"
- Open Paper Agent: Split-screen PDF + AI chat
- Ask Questions:
- "What is the main contribution?"
- "How does the attention mechanism work?"
- "Show me the key results"
- Smart Responses: AI answers with paper references
- Back to Canvas: Overview of research landscape
- Connection Mapping: See how papers relate
- Search Filter: Find specific topics instantly
- Future Vision: Mention scaling to thousands of papers
- Dark Theme: Professional research environment
- Gradients: Purple-to-blue for premium feel
- Glass Morphism: Subtle transparency and blur effects
- Micro-interactions: Hover states, loading animations
- Instant Feedback: No loading states over 1 second
- Smooth Animations: 60fps transitions
- Optimistic Updates: UI updates before API responses
- Progressive Enhancement: Works offline with demo data
- Search-First: Every action starts with a question
- Progressive Disclosure: Show complexity only when needed
- Visual Hierarchy: Clear information architecture
- Accessibility: Keyboard navigation, screen reader support
DelphiX/
├── app/
│ ├── components/ui/ # shadcn/ui components
│ ├── routes/
│ │ ├── _index.tsx # Research Hub (landing)
│ │ ├── knowledge-canvas.tsx # Interactive network view
│ │ └── paper.$paperId.tsx # Paper Agent interface
│ ├── tailwind.css # Global styles + animations
│ └── root.tsx # App shell
├── public/ # Static assets
├── package.json # Dependencies
├── tailwind.config.ts # Custom animations & theme
└── README.md # This file
- Visual Impact: Stunning UI that wows judges immediately
- Clear Value Prop: "NotebookLM for researchers who need deep connections"
- Interactive Demo: Judges can use it themselves
- Technical Excellence: Clean code, modern stack, smooth performance
- Scalability Story: Show how it handles thousands of papers
"Researchers spend 40% of their time just finding relevant papers.
DataEngineX changes that.
[Show landing page]
Instead of keyword search, you ask research questions.
[Type: 'How do neural networks learn representations?']
Our AI finds semantically relevant papers instantly.
[Create knowledge base]
But here's the magic - we don't just find papers,
we show you how they connect.
[Navigate to canvas]
This is your research landscape. Each paper is a node,
connections show relationships.
[Click on paper]
And when you want to go deep, meet your Paper Agent.
[Ask questions in chat]
It's like having a research assistant who has read
every paper and can explain any concept.
[Show canvas overview]
This is the future of research - not just search,
but understanding. Not just papers, but connections.
Not just information, but insights."
- Research Hub with search interface
- Knowledge Canvas with interactive nodes
- Paper Agent with AI chat
- Demo mode with mock data
- Modern UI with animations
- Backend API integration
- Real ArXiv search results
- User authentication
- Knowledge base persistence
- PDF upload and processing
- Collaborative knowledge bases
- Citation analysis
- Research trend prediction
- Export and sharing tools
- Mobile responsiveness
- Institution integrations
- Team workspace features
- Advanced analytics
- Plugin ecosystem
- ✅ Better Discovery: Semantic search vs basic upload
- ✅ Visual Connections: Network view vs linear interface
- ✅ Specialized for Research: Domain-specific features
- ✅ Open Ecosystem: ArXiv integration vs closed system
- ✅ AI-Native: Built for AI interaction from ground up
- ✅ Connection-Focused: Relationships, not just content
- ✅ Modern UX: 2024 interface standards
- ✅ Real-time: Instant feedback and updates
MIT License - Built for the research community
This is a hackathon project, but we welcome contributions for the open-source research community.
Built with ❤️ for researchers who want to understand how ideas connect