ResearchViewer

A web application for exploring and analyzing research papers from arXiv.

Live Site: researchviewer.org | API: researchviewer.org/api

Features

• Search & Discovery: Filter 2.9M papers by keywords, authors, topics, citations, and dates
• Topic Analysis: Explore 11K microtopics generated through clustering and semantic analysis
• Personal Library: Track reading lists, mark papers as read, and link your author profile
• Recommendations: Get personalized paper suggestions based on your reading history
• Analytics: View trends, citation patterns, and velocity metrics across fields

Tech Stack

Backend: Flask + Gunicorn, DuckDB (2.9M papers, 1.7M authors, 11K microtopics), Firebase Auth, Redis caching

Frontend: React + TypeScript, Vite, TailwindCSS, Recharts

Data: arXiv metadata enriched with OpenAlex citations, custom topic classifications, and semantic clustering

Database

Core Tables: papers (2.9M), authors (1.7M), microtopics (11K), paper_microtopics (5.2M)

User Tables: users, reading lists, read history, publications

Development

# Backend (runs on :8080)
python -m src.main

# Frontend (runs on :3000)
cd frontend && npm run dev

Deployment

Docker deployment with GitHub Actions CI/CD:

1. Push to main triggers build
2. Image pushed to GitHub Container Registry
3. Self-hosted runner pulls and restarts container