Skip to content
Neattrim

Transcribing interviews privately

· Editorial team

Do it now
Transcribe audio and video
Turn speech into text privately.

You can transcribe an interview without sending the recording to a server by using a tool that runs speech recognition in your browser. Open the file, pick the language, choose a format, and read the result. The audio stays on your device, so this suits confidential interviews, but you still need consent and a careful read-through.

Why privacy matters for interviews

Interview recordings often hold names, health details, opinions about employers and other things people said because they trusted you. Sending the audio to a third party adds a risk you may not be allowed to take. Journalists, researchers and HR teams often have rules about where recordings may go.

Before you choose a tool, check three things:

  1. Where does the audio go? The answer should be on a tool's page, not hidden in a policy.
  2. Is it kept? If the audio is uploaded, find out how long it is stored and who can open it.
  3. What did the participant agree to? Consent forms may name the tools you may use.

Step by step with an on-device tool

  1. Open Transcribe in a current desktop or laptop browser.
  2. Add the recording. Audio and video files are both accepted.
  3. Choose the spoken language, or leave it on automatic.
  4. Choose an output: plain text, text with timestamps, or a subtitle format such as SRT or VTT.
  5. Start. The first time, the speech model has to download, which takes longer than later runs.
  6. Read the result and fix mistakes before you quote anyone.

The tool page lists where each step runs and the file limits for your device. Long recordings may need to be split, and very large files may run better on a desktop than on a phone.

What affects accuracy

  • Audio quality. The biggest factor. A quiet room and a microphone near the speaker beat any setting.
  • Overlapping speech. Two people talking at once usually produces errors.
  • Accents and technical words. Names, places and specialist terms are the most likely mistakes.
  • Language. Set the language when you know it, especially for short clips.

If a recording is noisy, try the noise remover first and transcribe the cleaned file.

Check the text before you use it

A transcript from speech recognition is a draft. Do not quote it without listening to the part you quote.

  • Replay each quote and compare it with the text.
  • Check names, numbers and dates twice.
  • Mark the places where you were not sure, for example with a bracketed note, so you can check them later.
  • Timestamps help: use the timestamp format and you can jump to any line in the recording.

Who said what

A plain transcript does not say which person spoke. For interviews with two people, add the speaker names by hand as you read, or keep a separate note of who spoke when. Some modes can label speakers; check the tool page for what is available.

After you finish

  • Store the transcript and the recording in the place your project allows, and nowhere else.
  • Remove names that you promised to keep confidential before you share the text.
  • Delete the recording when your retention period ends.

FAQ

Is it really private?

When the tool runs on your device, the audio is processed in your browser and is not uploaded for transcription. The tool page explains where each step runs, and some options for long files can use servers, so read it before you start.

Recording and transcribing someone usually needs their consent, and the rules differ by country and by job. Follow your organisation's rules and any law that applies to you.

How accurate is it?

It depends on the recording. Clear speech in a quiet room gives good results, but you should still check every quote against the audio.

Can it handle a two-hour interview?

Device limits apply and depend on your computer. If a file is too long, split it into parts and transcribe each one.

Which format should I choose?

Choose plain text for reading, text with timestamps for checking quotes against the audio, and SRT or VTT if you also want subtitles for a video.

Related tools

Related guides