What's the best free interview transcription for journalists?

The one that fits the job's real constraint, which is confidentiality. A free tool that runs on your device and never uploads the audio protects your source better than a paid cloud service does. You give up speaker labels and timestamps, but you get every word for free with nothing leaving your laptop.

How do journalists transcribe interviews privately?

The private way is on-device transcription, where the model runs in your browser and the recording is processed locally. Nothing is uploaded, so there's no server copy of your source's voice to subpoena, leak, or repurpose. That's a stronger guarantee than a privacy policy, because there's no data on their side to make a promise about.

Does the transcript keep my exact quotes?

Yes. It returns a verbatim, word-for-word transcript, not a summary. That's what you want when a quote has to be exact and defensible. Read the transcript against the audio for the lines you plan to publish, since no automatic transcription is perfect on names, crosstalk, or heavy accents.

Does it label who said what?

No. There are no speaker labels and no timestamps. You get one clean block of every word spoken. For a two-person interview that's usually easy to follow, and you can mark the speakers yourself as you read it back.

Transcribe an Interview for Free (No Upload, for Journalists)

Q: Can I transcribe a one-hour interview for free?

Yes. Because the model runs on your machine, there's no server metering your minutes, so there's no per-minute cap and no paywall at the good part. The only real limit is your device's memory on very long files. For a two-hour sit-down, split the audio into two chunks and run them one after the other.

You finished the interview, and now you have an hour of audio to turn into text before you can quote a word of it. Most free transcription tools will do that fast, but almost all of them upload your recording to their servers first. When the audio is a source talking on the record, or off it, that upload is the whole problem. Here's how to get a free, word-for-word transcript with the recording never leaving your laptop.

Why uploading an interview is a real problem

Source confidentiality is the job. The moment you upload a recording, a copy of your source's voice lives on a company's servers, under their retention policy, their security, and their next decision about your data. A privacy policy is a promise, and a promise can be changed, breached, or handed over on request. For a public podcast that risk is fine. For a source who trusted you, it isn't.

Most "free" web transcribers are cloud tools. They send your audio out to run a big model somewhere else and send the text back. The upload rarely shows up on the button, so you can be shipping a sensitive interview off your machine without realizing you agreed to it.

How to transcribe an interview for free without uploading it

The honest fix is to use a tool that runs the model on your own device, in the browser, so the audio is processed locally and never sent anywhere.

Open an on-device transcriber in your browser. No account, no install.
Load your interview file, a voice memo, a phone recording, or a .wav or .m4a from a recorder. Or record straight into the page if you're capturing live.
The model runs locally on your machine and returns a word-for-word transcript.
Copy the text out and paste it into your notes or your doc.

The recording stayed on your laptop the whole time. There was nothing to upload, so there's nothing to trust a server with.

What "on-device" actually means, and how to prove it

On-device (also called in-browser) means the transcription model downloads to your browser once, then does the work using your own processor. Your audio is never transmitted. There's a simple test: after the first load, turn your Wi-Fi off. If it keeps transcribing offline, it's genuinely local. If it stalls without a connection, it was uploading all along. Run that test once and you'll know which kind of tool you're actually using.

Does it keep my exact quotes?

Yes. You get a verbatim transcript, every word, not a cleaned-up summary. That matters more for journalism than almost any other use. When you publish a quote, the phrasing has to be exactly what was said, and an AI note-taker that hands you a tidy paraphrase has thrown away the one thing you needed. Verbatim keeps the specific, the awkward, the on-the-record detail intact.

One honest note: no automatic transcription is perfect. Names, technical terms, crosstalk, and strong accents are where errors show up. Treat the transcript as a fast first pass, and for any line you're going to print, read it against the audio. That's normal practice for careful reporting whether the tool cost you nothing or a dollar a minute.

Can I transcribe a one-hour interview for free?

Yes, with no per-minute meter. Cloud tools cap free users because each transcription costs them compute. When the work happens on your device, it costs the tool nothing per use, so there's no minute limit and no wall that cuts off at the good part. Talk for as long as the interview ran.

The one real ceiling is your device's memory on very long files. A one-hour interview is fine on a normal laptop. For a two-hour recording, split it into two files and run them back to back. That's a memory constraint, not a paywall, so you never hit a "upgrade to continue" screen.

What it does not do (so you can plan around it)

Being straight about the limits saves you time. This kind of tool gives you one clean block of verbatim text. It does not add speaker labels, so it won't mark who said what. It does not add timestamps, and it does not export to SRT or Word. You copy plain text out and format it yourself.

For a standard two-person interview, an unlabeled block is easy to follow, and you can tag the speakers as you read it back for quotes. If your workflow depends on machine-generated speaker separation or timecoded captions, this isn't the tool for that piece of the job. It's the tool for getting every word, privately, for free.

Also worth stating plainly: on-device means the audio doesn't leave your machine, and that's the actual claim. It's not a legal or compliance certification. If your outlet has formal handling requirements for sensitive material, confirm those on your own terms.

That's the whole idea behind JustWrite: load the interview, get a clean word-for-word copy, free and private, because the transcription runs on your own device and your source's voice never leaves it.