Your Intelligent Research Companion
CORTEX is an AI-powered research assistant that helps users upload, analyze, and interact with academic papers. It combines seamless PDF management, retrieval-augmented generation (RAG), and multi-model chat capabilities within a clean, modern web interface.
CORTEX allows users to upload research papers, query them using natural language, and receive citation-linked answers grounded in their uploaded documents. The system is designed for students, researchers, and professionals who want to save time digesting and comparing academic sources.
- Upload multiple research papers (PDFs) and have them automatically parsed and indexed.
- Ask natural language questions across your entire document library.
- Get answers supported by direct citations and page numbers.
- Chat conversationally with CORTEX about your uploaded documents.
- Switch between different large language models (e.g., GPT-4, Claude, Gemini).
- Citations in responses are clickable and jump directly to the referenced PDF page.
- Each upload session forms a workspace, grouping related documents.
- Every workspace maintains its own vector index and context memory.
- Secure JWT-based login and signup.
- Each user’s uploads, workspaces, and chat history are fully isolated.
- Built-in React-PDF viewer with highlighting support.
- View, scroll, and reference PDFs directly in the app without downloading.
- Click on citations to navigate to the exact page mentioned.
| Layer | Technology |
|---|---|
| Frontend | Next.js 14 (App Router), TypeScript, TailwindCSS |
| Backend | FastAPI (Python) |
| Database | MongoDB Atlas |
| Authentication | JWT tokens (Cognito-ready) |
| File Storage | AWS S3 (planned) |
| AI Models | OpenAI, Anthropic, Gemini APIs |
| Deployment | Railway (FastAPI) + Vercel or Netlify (Next.js) |
CORTEX/
│
├── frontend/ # Next.js app
│ ├── components/ # Chatbot, PDF viewer, navigation, etc.
│ ├── app/ # Pages and routes
│ ├── public/ # Static assets
│ └── utils/ # API calls, helpers
│
├── backend/
│ ├── src/
│ │ ├── routes/ # Upload, papers, authentication
│ │ ├── main.py # FastAPI entry point
│ │ └── models/ # Pydantic schemas and database models
│ └── requirements.txt
│
└── README.md
- Fine-tuned document summarization and section retrieval
- Support for highlighting answers directly in PDFs
- S3 integration for persistent file storage
- User collaboration and shared workspaces
This project is licensed under the MIT License.