How to Transcribe an Interview or Meeting Without Typing It All

Why not start from scratch?

Transcribing by hand can take several times longer than the audio itself. A one-hour interview could take an entire afternoon. Automatic transcription changes the job: instead of typing everything, you review a transcript that’s already close to finished. It still takes care, but it saves a lot of time.

This method works for research interviews, work meetings, classes, and testimonials.

Step 1: Prepare your audio

• File format: The transcription tool supports common audio and video formats. If your file uses an unusual format or you get an error, convert it to MP3 with Convert audio.
• Cut out what you don’t need: Remove the conversation before you start, breaks, and goodbyes. Use Cut audio to set the start and end points.
• Split very long recordings: On phones and less powerful computers, transcribing 20- to 30-minute sections is more reliable than processing a two-hour file. You can also review one section while the next is processing.

If you haven’t recorded yet, these tips on how to record high-quality audio on your phone will make a bigger difference to your transcript than any adjustments you make later.

Step 2: Transcribe your audio

1. Open Transcribe audio to text and select your file.
2. Choose an accuracy setting: Fast for a quick test or on phones, Balanced for most recordings, or High on a computer for difficult audio. The first time you use a model, it downloads and is saved in your browser.
3. Select the spoken language: Portuguese, English, or Spanish.
4. Choose to generate a transcript and subtitles (.srt). Subtitles include timestamps for each segment, making it easier to find a line in the audio as you review.
5. Start the transcription and wait. Processing time depends on your device and the audio length.

All processing happens in your browser. The interview audio isn’t uploaded to any servers, which matters when the content is confidential or includes research participant data.

Step 3: Review while listening

Automatic transcription is most likely to get names, acronyms, technical terms, numbers, and overlapping speech wrong. To review efficiently:

• Listen at 1.25x or 1.5x speed while following along with the text, and pause only when you spot an error.
• Use the SRT timestamps to jump straight to sections you’re unsure about.
• Mark anything you can’t understand as [inaudible] instead of guessing.
• Check the punctuation carefully—it can change the meaning of what was said.

Step 4: Identify the speakers

The tool doesn’t automatically separate speakers. During your review, add a label at the start of each person’s turn, such as I: for the interviewer and P1:, P2: for participants. For meetings, use each person’s initials. It helps to note the order in which people speak while recording.

Transcribing interviews for academic research

• Verbatim or edited: Ask your advisor whether the transcript should include hesitations, repetitions, and filler words (like “you know” or “like”), or whether you can edit it for readability.
• Anonymization: Replace participants’ names, companies, and locations with codes, following the consent form and your ethics committee’s guidelines.
• Store the original recordings securely for as long as your project requires.

For work meetings, the full transcript is rarely the final deliverable. After reviewing it, turn it into concise meeting minutes that capture decisions, owners, and deadlines. Keep the full transcript for reference, and share the minutes.

Frequently asked questions