How do I transcribe a podcast episode for free?

Open an on-device transcriber in your browser, load the episode's audio file, and let the model run locally. You get a word-for-word transcript back with no account and no per-minute charge. The file stays on your computer the whole time.

What's a good transcript tool for podcasters?

For show notes and a searchable transcript, you want something that gives you the verbatim text and doesn't cap you or upload your file. On-device tools fit that job well. If you need speaker labels or timestamped subtitles, you'll want a different tool, since a plain on-device transcriber gives you the text only.

Can I transcribe a long episode without hitting a cap?

With a true on-device tool there's no server metering your minutes, so there's no per-minute cap to run into. The real limit on a very long episode is your browser's memory, not a paywall. If a two-hour file struggles, split it into a couple of chunks and run them separately.

Does it upload my audio file?

No, if the tool runs the model on your device. The audio is processed locally in the browser and never sent to a server. You can test it: load a file, turn off your Wi-Fi, and if it keeps transcribing, nothing was being uploaded.

How accurate is the transcript?

It's built on Whisper, the same model family behind many paid cloud services, so for clear speech the accuracy is close to what those charge for. Crosstalk, music beds, and heavy background noise lower it, same as any tool. You'll usually do a quick cleanup pass before publishing, but the bones are there.

How to Transcribe a Podcast Episode to Text for Free (No Upload)

You finished an episode and now you want the transcript, for show notes, for a blog version, or just so people can search what you said. Here's the short answer: you can transcribe a podcast episode to text for free, in your browser, without uploading the file anywhere. Load the audio into an on-device transcriber, let the model run locally on your own machine, and copy out the word-for-word text. No account, no per-minute meter, and the file never leaves your computer.

That last part is the part most tools skip over, so let's walk through how it actually works.

How to transcribe a podcast episode for free

The flow is short:

Open an on-device transcriber in your browser. Nothing to install, nothing to sign up for.
Load your episode's audio file (an MP3 or WAV export from your recording setup works fine).
The model runs locally and returns a verbatim transcript of the whole episode.
Copy the text out and paste it wherever you're building show notes or a post.

The key detail is step three. The transcription happens on your device, using your own processor, so there's nothing to send off and nothing to wait on a server for. It's the same idea as running the model yourself, without the command line and the setup.

Why most free podcast transcribers upload your file

Transcription takes real compute, and the easy way to build a tool is to push your file up to a server, run a big model there, and send the text back. That's what most free web transcribers do, including several that market themselves as private. Your episode ends up sitting on their infrastructure, under their retention policy and whatever they decide to do with recordings next.

For a public episode you're about to publish anyway, that's not a disaster. But it's still your raw file, sometimes with the cold open, the bloopers, and the bits you cut. And plenty of these "free" tools cap you at a handful of minutes or files a day, then ask for a card once you hit the wall. On-device sidesteps both problems: there's no upload, and there's no meter, because each transcription costs the tool nothing to run.

Is there really no cap on a long episode?

There's no per-minute cap, no. Because nothing is being metered on a server, a 90-minute episode costs the same as a 5-minute one: nothing. That's the honest advantage over the free tiers that stop you at 30 or 120 minutes a month.

The one real limit is your browser's memory on very long files. A short episode is no trouble. If you're transcribing a two-hour marathon and your machine chugs, split the audio into two or three parts and run them back to back, then stitch the text together. It's a couple of extra minutes, and you're still not uploading anything or paying anyone.

What you get back, and what you don't

You get a verbatim transcript: every word, in order, the way you said it. That's exactly what you want for show notes, a searchable archive, or a blog version of the episode, because the specific phrasing is usually the point. A summary would throw away the good lines.

Be honest with yourself about the format, though. A plain on-device transcriber gives you the text and that's it. It won't label which co-host said which line, and it won't hand you a timestamped subtitle file. If your workflow needs speaker names or an SRT for captions, that's a different kind of tool. For turning talk into clean, copyable text, this is the simple, free lane.

Does the transcript need cleanup?

Usually a little. On-device transcription is built on Whisper, the same model family behind a lot of paid cloud services, so for clear speech in a decent recording the accuracy holds up well. Where it slips is the stuff that trips up every tool: crosstalk when two people talk over each other, music beds under the intro, proper nouns and brand names, and heavy background noise.

So expect to skim it and fix a few words before you publish, especially names. Auto-punctuation from Whisper is decent, so you won't be starting from a wall of text. For a fast pass, a good on-device tool often includes an optional cleanup step that tidies the raw output, though that one step does need a connection to run.

Getting the text out for show notes

Once the transcript is done, copy it and paste it into wherever you keep show notes: your notes app, your CMS, a doc. From there you can pull out the timestamps you care about by hand, chop it into chapters, or feed the whole thing to an AI tool to draft a summary or a set of episode chapters. The verbatim text is the raw material; what you build from it is up to you.

Because the file never got uploaded, there's no cloud copy to clean up afterward and no export paywall standing between you and your own words.

That's the whole idea behind JustWrite: load your episode, get a clean word-for-word copy, free and with no cap, because the transcription runs on your own device and the audio never leaves it.