DistillNote
DistillNote — Structured notes from YouTube/podcasts in 60s
Hey IH! After 8 weeks of building, I just launched DistillNote.
The problem: We all consume tons of content but retain almost none of it. Bookmarks pile up. AI summarizers give you a paragraph. Neither is a system.
The solution: Paste a URL -> get organized notes (TL;DR, chapters, takeaways, timestamps, action items) -> everything goes into a searchable vault where you can ask questions and get cited answers.
Business model:
- Free: 30 credits/month (1 credit = 1 minute of content)
- Plus: EUR 7.99/mo (500 credits + semantic search + Q&A + clean exports)
- Pro: EUR 19.99/mo (1,500 credits + priority processing + model choice)
Revenue so far: Pre-revenue. Just launched this week.
Cost structure: Gemini 2.5 Flash is the main cost (~$0.003 per minute of content). Groq Whisper for audio transcription. Hosting on Vercel free tier + Neon free tier + Railway (~$5/mo for the worker). Total infra cost at low scale: under $20/mo.
What's next: Article ingestion (any URL), podcast RSS support, browser extension.
Would love your feedback on the product and the pricing.
3 Comments
Turning long-form content into structured notes in seconds is definitely something a lot of people need right now.
Are you focusing more on students, creators, or general productivity users initially?
Honestly, all three! But the people who get hooked fastest are students tired of re-watching lectures, creators drowning in podcast content, and knowledge workers who watch conference talks and forget everything a week later.
The core pain is the same across all of them - you consume something valuable and it just... disappears. That's what we're fixing.
What's your use case?
My use case would probably be more around long-form content and research — podcasts, talks, articles, and turning them into structured notes or summaries I can come back to later.
A lot of good information gets consumed once and then forgotten, so something that turns content into searchable, structured knowledge would be very useful.