Transcribe.soTranscribe.so

Turn YouTube videos, lectures, and podcasts into notes you can search and ask.

Paste a YouTube link, drop in a recorded lecture, or upload a podcast. Every transcript joins one searchable library you can question from the web, ChatGPT, or Claude. Ask where the professor defined a term, where a guest cited a study, or where a concept first came up. Get a cited answer with a timestamp and jump to the exact second they said it. Works on Korean MOOCs, Japanese podcasts, Spanish talks, and English lectures.

No credit card required.·Pay only for what you use.

See it in action

Real output from a real transcription

Browse chapters, ask questions, and explore search results from an actual transcript.

44 Harsh Truths About The Game Of Life - Naval Ravikant (4K)
Chris Williamson
Contents
8 chapters · 513 sections
1Happiness Versus Success: Philosophical Reflections on Contentment, Desire, and Motivation
2Optimizing Sleep: Smart Temperature Regulation and the Foundations of Self-Esteem
3Decisive Action and Iterative Practice: Keys to Optimal Choices and Mastery
4Wealth Management: From Materialism to Value Creation and Fair Compensation
5Evaluating LLMs: Capabilities, Limitations, and Their Role in AI's Evolving Landscape
6Pathogens, Evolution, and Knowledge: How Humans Adapt and Defend
7Agency, Power, and the Individual: From Child Development to Cultural Conflict
8Unseen Trends: Media Oversights, Medical Limitations, and the Primitive State of Modern Biology
Q&A preview
Answer
Naval explains two distinct paths to happiness using the story of Alexander and Diogenes. The first path is through success—conquering the world, satisfying material needs, and getting what you want. The second path, exemplified by Diogenes living in a barrel, is simply not wanting in the first place. As Socrates said when shown luxuries: 'How many things there are in this world that I do not want.' Naval suggests not wanting something is as good as having it—both paths lead to the same destination of contentment [00:38–01:10]. He's not sure which path is more valid, noting it depends on how you define success [01:10–01:25].

Command Palette

Search for a command to run...

You remember the answer is in there. You just can't find it.

  • You scrub a 3-hour lecture to relocate one explanation and overshoot every time
  • AI summaries throw away the exact phrasing you actually need to study from
  • One-shot transcripts pile up. There is no search across last week's lectures and last month's podcasts
  • Single-engine tools mangle non-English audio. Korean MOOCs and Japanese podcasts come back broken

What you get with a searchable library of everything you study

One searchable library across every lecture and podcast

Every YouTube link, recorded class, and podcast episode joins one library. Search by meaning, not just keywords, across hours of content you have studied.

Ask anything. Jump to the second they said it.

Ask 'where did they define entropy?' or 'which episode covered the 2008 crash?' Get a cited answer with a timestamp. Click the citation to land in playback at the exact moment.

Chapters, topics, and takeaways on every transcript

Long lectures and podcasts come back with a chapter spine, per-section topics, and a takeaways block. Jump between sections instead of scrubbing a flat timeline.

Works in any language, automatically

67 languages with measured accuracy per language. Korean MOOCs, Japanese podcasts, Spanish talks, Portuguese lectures, English seminars. The right engine is picked for you, so you can study instead of comparing models.

Notes ready for Notion, Obsidian, and Apple Notes

Copy a clean transcript, a chapter outline, or a takeaways block as markdown. Paste into Notion, Obsidian, Apple Notes, or share via Slack and WhatsApp without reformatting.

Long-form audio is the default, not a stretch goal

3-hour lectures, Lex Fridman episodes, full Stanford CS courses, multi-hour Karpathy walkthroughs. Long audio is what the pipeline is tuned for.

Ask your library from ChatGPT or Claude

Add the public ChatGPT GPT or the Claude Custom Connector and you can quiz your study library from inside the AI chat you already use. Same wallet, same per-minute pricing, no extra account.

What people use this for

  • Paste a recorded lecture and ask 'where did they define X?' with a cited timestamp
  • Turn a YouTube course playlist into a searchable library of structured notes
  • Find the exact moment a 3-hour podcast covered a specific study or guest name
  • Study Korean MOOCs, Japanese podcasts, or Spanish talks with accurate non-English transcripts
  • Copy chapter outlines and takeaways into Notion, Obsidian, or Apple Notes
  • Search across last semester's lectures to recall a definition without rewatching anything
  • Quiz your study library from ChatGPT or Claude without leaving the chat
  • Capture lectures and study sessions live on a Mac, then study from the transcript

FAQ

Frequently asked questions

Generate a structured transcript with chapters and a cited Q&A layer, not a one-paragraph summary. Paste a YouTube link, get back a transcript with chapters, topics, and a takeaways block. Ask a question and the answer comes back with a timestamp you can click to jump straight to the moment in playback.

Yes. Beyond Q&A on a single transcript, you can ask and search across every transcript in your account. Find a definition, a guest name, or a study citation across last semester's lectures and last month's podcasts, even if you can't remember which video it was in.

Yes. Paste a YouTube link or upload audio and video files in 67 supported languages. Korean university MOOCs, Japanese podcasts, Spanish talks, Portuguese lectures, and English seminars all flow through the same pipeline. The right speech-to-text engine is picked per language with measured accuracy.

67 languages with FLEURS-measured accuracy per language and per pipeline. A Korean MOOC, a Japanese podcast, and an English lecture each route to the engine that handles them best. You don't pick a model. The defaults are tuned for long-form learner audio.

Yes. Every Q&A answer cites the segments it drew from. Click a citation and playback jumps to the exact second. This is the trust unlock for studying from AI answers: you can verify the source in one click instead of hoping the summary got it right.

Free trial credits on signup with no credit card required. Enough to transcribe roughly fifteen minutes on the default engine, so you can test your own audio before paying. After that, pay-per-minute with no monthly minimums.

You don't have to. The default routes each file to the engine that handles it best for the language and content type. Power users can override the model from inside the app if a specific pipeline (Qwen3-ASR-Flash, GPT-4o Transcribe, Voxtral, etc.) is preferred.

Want a deeper comparison? Best YouTube-to-notes tool for learners

Build a library from one YouTube link.

Paste any lecture, podcast, or video. Get chapters, takeaways, and a cited Q&A layer in one pass. Free credits to start. No card required.