DropLine: An Asynchronous RAG Architecture for Resolving Wrapper Link Bottlenecks
DropLine is an asynchronous RAG-based system that resolves wrapper and redirected web links and transforms the actual content into structured educational material. The system handles URL resolution, content extraction from webpages and videos, text cleaning, and LLM-based generation of summaries, key concepts, explanations, applications, and quizzes. It is implemented using FastAPI, Trafilatura, youtube-transcript-api, Gemini 2.5 Flash, and Streamlit, with session-state support for follow-up questions.
Authors
- Abhijit Yamgar
- Aditya Khorne (ORCID: https://orcid.org/0009-0000-8534-2832)
- Ajay Virkar
Institutions
- Savitribai Phule Pune University (IN)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-04-30
- DOI
- https://doi.org/10.5281/zenodo.19923632
- Primary Topic
- Web Data Mining and Analysis
- Type
- article
- Field-Weighted Citation Impact
- 0.00