Show HN: ekoAcademic – Convert ArXiv papers to interactive podcasts (www.wadamczyk.io)

ekoAcademic is a lightweight tool built by Aidan McConnel, Shaan Amin and collaborator that converts newly released arXiv papers into short, accessible audio summaries and interactive podcasts. The system mirrors arXiv categories, extracts and summarizes new papers, and produces cached non‑interactive audio clips so once a paper’s summary is generated it needn’t be re-created for other listeners. Crucially, it pairs those clips with GPT-driven real‑time sessions for conversational Q&A and multi‑paper summarization: users can grant microphone access, interrupt playback, and ask follow‑up questions in many languages, with responses returned in the same language. The team says the pipeline is cheap enough for quasi‑real‑time generation and could be extended beyond arXiv to other databases. For the AI/ML community this lowers the friction to scan and digest new literature during commutes or chores, making triage and discovery faster and more accessible—especially for multilingual audiences or those with visual impairments. Technical implications include reliance on LLMs for summarization and QA (so accuracy and context fidelity are potential concerns), a caching model that reduces compute cost, and the possibility to index and summarize sets of papers on demand. The project seeks community feedback on subject coverage, translation needs, accuracy tradeoffs and feature requests if others find interactive audio literature useful.