Free podcast transcript generator
Full-episode transcripts with timestamped chapters and per-chapter summaries — the raw material for show notes, clips and SEO pages. Free for unlimited episodes, because it runs in your browser.
Free plan today: WAV files of any length; MP3/M4A up to 1 hour per file for now. For a 2-hour episode, export WAV from your editor (every DAW and editor can), split the MP3, or use Pro cloud mode.
Transcript → chapters → show notes
A transcript on its own isn't show notes — it's 10,000 words of unpunctuated conversation. Record to Notes structures it while it transcribes: chapters are detected at real topic changes with timestamps, and each chapter gets a summary. Those chapter summaries are already the skeleton of your show notes and your episode page; the SRT export is your caption file; the timestamps are your clip markers.
Drop the episode
The exported audio file from your editor or hosting platform. Video episodes work too — the audio track is used.
Get the transcript live
Streams in as it processes, so you can start pulling quotes before the file finishes. Chapters land at real topic changes.
Publish more
SRT for captions, Markdown for your site, TXT for everything else. Pro drafts show notes and articles for you.
Does it label who's speaking?
Not yet — being straight about this: Record to Notes gives you the full transcript with timestamps and auto-detected chapters, but it does not currently label speakers. For a multi-host show where per-speaker labels are non-negotiable, this is the one thing the tool won't do for you today. Speaker labels are on the roadmap; everything else on this page works now.
What a back catalog actually costs to transcribe
Most transcription tools run on their servers and price accordingly — fine for one episode, painful for a hundred. On-device transcription has no per-minute cost, which is what makes transcribing your entire archive plausible.
| Tool | Free plan limits | A 100-episode back catalog |
|---|---|---|
| Record to Notes | No episode cap, no signup (WAV any length; MP3/M4A ≤1h per file) | All of it, free — queue them one by one, or batch with Pro |
| Otter.ai | ~300 min/month, 30 min per conversation | ~5 hour-long episodes per month, in 30-min chunks |
| TurboScribe | 3 files per day, 30 min each | Hour-long episodes don't fit in one file |
| Notta | Monthly minute quota with short per-file caps | A handful of episodes |
Competitor limits as of September 2026 — plans change, check their pricing pages. The economics don't: server-side tools pay per minute of your audio; your browser doesn't.
The details producers ask about
- Formats: WAV at any length; MP3/M4A up to 1 hour per file for now. Video files are accepted — the audio track is extracted.
- Speed: roughly 10 minutes for a 2-hour file locally with GPU acceleration; about a minute with Pro cloud mode.
- Privacy: on the free plan the episode is processed entirely in your browser — zero network calls during transcription. Unreleased audio stays unreleased.
- Output: transcript with timestamps, chapters with summaries, exported as SRT, TXT or Markdown. The Markdown pastes cleanly into your CMS or Notion.
- Languages: the local models handle English best; Pro's cloud model adds high accuracy across 90+ languages.
Pro — for shows on a schedule
- Batch queue: drop the whole back catalog folder and let it run
- Cloud speed: a 2-hour episode transcribed in about a minute — publish-day friendly
- Highest accuracy in 90+ languages — crosstalk, guests on bad mics, technical topics
- Show-notes drafting: structured notes and article drafts from the transcript
FAQ
Is there an episode length limit?
WAV files: no limit — 3-hour episodes are fine. MP3/M4A: capped at 1 hour per file for now, so export WAV from your editor for long episodes, or use Pro. Processing runs on your device, so nobody is paying per minute.
Why is it free? What's the catch?
Your browser does the transcription on your own machine, so we don't pay a cloud bill for your audio — and don't need to cap you or charge you to cover one. The paid Pro tier exists for people who want cloud speed, higher accuracy and batch processing. That's the whole model.
Does it label who's speaking?
Not yet. You get the full transcript with timestamps and chapters, but speakers aren't labeled. If per-speaker labels are essential for your multi-host show, this tool won't replace that part of your workflow today — it's on the roadmap.
Do you upload my unreleased episode?
Not on the free plan — audio is processed locally in your browser and never leaves your computer. That matters when the episode isn't out yet.
Can I get SRT captions?
Yes — export SRT with timestamps for captions, or TXT/Markdown with chapters included for show notes and episode pages.
How accurate is the transcript?
Clean, well-mic'd episodes come out very usable on the local Small model. Crosstalk, remote guests on laptop mics, or heavy accents are where Pro's cloud model earns its keep.
Can I edit the transcript before publishing?
Export TXT or Markdown and edit in anything — your CMS, Notion, Google Docs. Timestamps survive the export, so caption timing stays intact in the SRT.