How to transcribe an interview from recording to text in minutes.
To transcribe an interview, upload the recording (mp3, m4a, wav, mp4 or mov) and pick the language spoken. Braindump turns it into a transcript split by speaker, with a start time on every turn and a short summary. Rename Speaker 1 and Speaker 2 to Interviewer and Participant, check it against the audio, then export to Word, PDF or TXT. 3 free transcriptions a day, no credit card.
Interview transcription in 3 steps
Upload the recording
Recorded on your phone, a voice recorder, Zoom or Teams? Upload the file as it is: mp3, m4a, wav, mp4, mov and more. Free files can be up to 2 hours long, Pro files up to 5 hours.
Pick the language spoken
Choose the interview language from 59 options. Braindump doesn't guess the language, so you pick it. If the interview is mostly in English with a few terms from another language, pick English.
Get the transcript and summary
Within minutes the full transcript is in your dashboard with a title and a short summary. When two or more voices are found, it comes as speaker turns, each with the time it starts. You can close the tab: we email you when it's ready.
From transcript to analysis
A transcript is where the analysis starts: the text you check, quote, code and attach to your dissertation.
Name the speakers, then check the audio
Braindump labels each speaker as Speaker 1, Speaker 2. Rename them Interviewer and Participant (or P1, a pseudonym) once and every turn updates. Then play the recording back, fix names and technical terms, and move any turn that landed on the wrong speaker.
Export to Word, PDF or TXT
Download each transcript as DOCX, PDF or TXT, with the speaker names you set at the start of each turn. Import the DOCX or TXT into your analysis tool, such as NVivo or ATLAS.ti, or add the PDF to your dissertation appendix.
Find any quote in seconds
The full text is searchable. Find the passage you want to quote, or every mention of a name you need to swap for a pseudonym, without scrubbing through the audio.
A summary for every interview
Most projects have 8 to 20 interviews. The short summary on each one shows what that interview covered at a glance, so you know which ones to reread when you build your themes.
Example: an interview transcript with speaker labels
A made-up excerpt, not a real participant, showing the shape of the result after Speaker 1 and Speaker 2 were renamed.
Night shifts and study time
The participant keeps Mondays for errands because night shifts leave her too tired during the week, and studies mostly at weekends.
Interviewer00:00
How do you organise your week when you're on night shifts?
Participant00:06
I try to keep Mondays for shopping and things at home, because the rest of the week I'm just too tired.
Interviewer00:19
And does that affect when you study?
Participant00:23
Quite a lot. I mostly study at weekends, and some weeks not even then.
In the DOCX, PDF and TXT export each turn starts with the name, for example "Participant: I try to keep Mondays...". The start times stay in your dashboard.
Transcribing qualitative interviews for a dissertation
Typing up an interview by hand means play, pause, rewind, type. A common rule of thumb is four to six hours of typing for every hour of audio, so 12 semi-structured interviews of 45 minutes each can take 36 to 54 hours before you code a single line. With Braindump the AI writes the first draft, already split into interviewer and participant turns, and you spend that time checking it instead. You get readable turns of what was said, which is what most thematic analysis works from. If your method calls for strict verbatim, with pauses and overlaps marked, treat the AI text as a draft and add those details while you listen back. Either way, check your consent forms and your ethics board's data rules before you upload participant recordings.
Not a good fit if you need word-level timestamps, subtitle files, Jefferson-style notation of pauses and overlaps, or live transcription during the interview. Large focus groups recorded on one phone also need more manual fixing of speaker turns.
Questions about transcribing interviews
How do I transcribe an interview for free?
Sign in with Google or Apple, upload the recording and pick the language spoken. You get 3 free transcriptions a day, no credit card needed. Free files can be up to 2 hours long, but for recordings over 30 minutes you see a preview of the transcript and the full text needs Pro.
How long can an interview recording be?
With Pro, files can be up to 5 hours or 600 MB each, with fair use. On the free plan files can be up to 2 hours: interviews over 30 minutes show a preview of the transcript, and the full text needs Pro.
Does it label who is speaking (speaker diarization)?
Yes. Braindump labels each speaker automatically, on the free plan and on Pro, with nothing to switch on. When it finds two or more voices, the transcript comes as turns marked Speaker 1, Speaker 2, and a card shows each person's talk time. Rename them Interviewer and Participant, move a turn to the other speaker, or merge two speakers, with undo. Speaker labels are on the web for now, not yet in the iOS and Android apps.
Does the transcript include timestamps?
Each speaker turn shows the time it starts, such as 12:40, so you can find a quote in the recording. There are no word-level timestamps and no subtitle files, the start times stay in the dashboard rather than in the export, and a recording with only one voice comes back as plain paragraphs without times.
Does it work for focus groups?
It works best with 2 or 3 clear voices. With 4 or more people around one phone on a table, voices can get mixed up, so check the turns against the audio and move or merge speakers where needed. For Chinese, Galician, Azerbaijani, Swahili, Welsh, Icelandic and Maori, files over about 23 minutes come back without speaker labels.
Can I import the transcript into NVivo or ATLAS.ti?
Download it as DOCX or TXT and import that file into your analysis tool. Each turn starts with the speaker name you set, such as Interviewer: or P3:. You can also export a PDF for your appendix or copy the text straight into your notes.
Is an AI transcript accurate enough for a dissertation?
It gives you a strong first draft, not a finished transcript. Clear recordings in a quiet room come out best. Always check the text against the audio before you quote or code it, which is good practice with any transcript.
Which languages and file types does it support?
59 languages, including English, Spanish, French, German, Arabic, Hindi and Japanese. It doesn't detect the language, so choose it before you upload. Files can be mp3, m4a, wav, aac, mp4, mov and more, so Zoom and Teams recordings work as downloaded.
How long does it take?
Usually a few minutes, even for an hour-long interview. You don't have to keep the page open: we email you when your transcript is ready.
Can I use it for confidential research interviews?
Your recording is used to create your transcript. Before you upload participant interviews, check what your consent forms, your ethics board and your university's data rules say about transcription services, and replace names with pseudonyms if your protocol asks for it.
Interviews done? Let us do the typing
3 free transcriptions a day, no credit card needed.
Transcribe your first interviewTranscribing lectures, meetings or voice memos too? Use the same audio to text tool for any recording. Longer interviews and heavier projects: see Braindump pricing.