Podcast to text: transcribe every episode

A podcast transcript is how an episode gets found. Search engines can’t index audio, but they can index the text under it — and that same text becomes your show notes, your pull quotes, your captions, and the version deaf and hard-of-hearing listeners can actually read. The reason most shows skip it is cost: per-minute transcription services turn a back catalogue into a real bill.

This tool removes the meter. Episodes are transcribed in your browser, on your own machine, so there is no per-minute charge and no upload — a single episode or an entire season’s backlog is just runtime. Your files are never uploaded — everything runs on your device. An unreleased or embargoed episode can be transcribed the day you finish the edit, without it ever leaving your computer.

Checking what this browser supports…

How to transcribe a podcast episode

  1. Drop in the episode

    Add the exported audio — MP3, WAV, M4A, FLAC — or the raw video file. There is no minute cap and no queue, so length doesn’t matter.

  2. Watch the transcript stream in

    Timestamped segments appear while the file processes, with progress and an ETA visible the whole way through an hour-long episode.

  3. Export the format you need

    TXT for show notes, SRT or VTT for captions, JSON to feed your own pipeline. Fix any misheard names inline before you export.

Built for publishing episodes

Findable episodes

Publishing the transcript on the episode page gives search engines the words your audio hides — every name, product, and topic becomes something a listener can find you by.

Show notes, the easy way

The full transcript is the raw material for show notes, chapter summaries, and pull quotes. Draft from text you already have instead of re-listening at 1x.

A whole backlog, no meter

With no per-minute cost, transcribing a hundred back-catalogue episodes costs only the time they take to run — and nothing you haven’t released yet is ever uploaded.

Podcast transcription questions

Can I publish the transcript on my episode page?
Yes — that is one of the best reasons to make one. The exported TXT drops straight into a show-notes field or a webpage, giving search engines indexable text for audio they otherwise can’t read.
Does it separate the host from the guest?
Not yet. The transcript is a single timed stream without speaker labels. Timestamps make it easy to mark who’s talking while you edit — an honest limitation that’s on the roadmap.
Can I make captions or subtitles from it?
Yes. Export SRT or VTT and upload it alongside a video version of the episode, or use it as a starting point for a captions file on YouTube or your host.
Can it handle my entire back catalogue?
There’s no episode limit and no per-minute charge, so a full season or a multi-year backlog is only a question of runtime. Run them one after another and let it work.
Is it safe for an unreleased episode?
The episode is transcribed on your machine and never uploaded, so an embargoed or pre-release cut stays entirely on your computer. (This describes how the tool works, not legal advice.)
What powers the transcription, and what does it cost?
OpenAI’s open-source Whisper model, running locally in your browser via our open-source engine — WebGPU when available, WebAssembly otherwise. It’s free, with no account and no upload.