I’m Noah and I built ResearchPod because I was drowning in arXiv PDFs but had a 45 minute commute where my brain was free but my eyes were busy.
ResearchPod transforms any research paper (via PDF upload or arXiv browse) into a podcast episode that's about 20 minutes long and features two hosts.
How it works:
1. Extraction: It parses the PDF and cleans the text. 2. Analysis/Scripting: It uses OpenAI and Grok to extract key insights and reformat them into a natural dialogue between two hosts, Alex and Sam. They don't just read the text. Instead, they discuss the methodology and implications like peers. 3. Synthesis: The script is sent to a dedicated TTS service running Gemini 2.5-flash-preview. I'm using the multi-speaker features to give each host a distinct and natural voice.
Why I built this: I found that triage is the hardest part of research. I don't always have time to sit and read 30 pages just to find out if a paper is relevant to my interests. Listening to a high-fidelity summary while I’m at the gym or commuting helps me decide which papers deserve a deep dive later.
Tech Stack:
App: Swift (iOS 16+) Backend: Node.js on Vercel Database: Supabase TTS: Railway (handling the long-duration audio generation) The app is currently free to use. You can generate up to 3 episodes per day for free and you can listen to as many as you want.
I'm really looking for feedback from the research community here. Does this bridge the gap for you or is the conversational format too "lite" for your workflow?
App Store: https://apps.apple.com/de/app/researchpod/id6751007088?l=en-GB
Web Share (to hear an example): https://researchpod-share.vercel.app/episode/84435c8e-b691-4f58-82bd-f09a6e281d11
Would love to hear your thoughts!