...It was created to make the large daily flow of new submissions easier for researchers to manage. The backend downloads papers, extracts their text, creates thumbnails, and computes TF-IDF representations for similarity analysis. Users can search papers, find work similar to a selected paper, browse popular submissions, and maintain a personal library. Personalized recommendations are generated through models trained from user preferences. The web interface uses Flask, Tornado, and SQLite, while the indexing pipeline relies on scientific Python libraries. ...