Interview transcript template
Header, speaker labels, timestamps and formatting conventions.
A transcript is a document other people will quote from, so its formatting is not decoration. Consistent speaker labels and honest markers for what couldn't be heard are what make a transcript usable as evidence, as research data, or as source material for a piece of writing.
The transcript template
INTERVIEW TRANSCRIPT Project / publication: [name] Interviewee: [name, role, organisation] Interviewer: [name] Date and place: [date, location or platform] Recording length: [hh:mm:ss] Transcription type: [verbatim / clean verbatim / edited] Transcribed by: [name or tool, date] Consent: [recorded consent at hh:mm:ss / written consent on file] --- [00:00:00] INTERVIEWER: [question] [00:00:14] SURNAME: [answer] [00:01:02] INTERVIEWER: [question] [00:01:20] SURNAME: [answer] --- Notes: [anything relevant that the audio doesn't capture, a document shown, an interruption, a change of room]
Verbatim, clean verbatim or edited
Verbatim keeps every "um", repetition and false start. Required in legal and clinical work, and in research where hesitation is data.
Clean verbatim removes fillers and stutters but changes no words. The default for journalism and most research.
Edited tidies grammar so quotes read well. Acceptable for publication if you say so, and never acceptable when the exact phrasing is the point.
Decide before you start and state it in the header. A transcript that switches style halfway is worse than either.
Conventions worth following
[inaudible 00:12:31]: mark it with a timestamp rather than guessing.[unclear: sounds like "framework"]: a guess, labelled as one.[crosstalk]: where speakers overlap and no clean reading exists.[laughs],[pauses],[long pause]: only where they change the meaning.[Name]: for anything you've anonymised, kept consistent throughout.- Speaker labels stay identical every time: the same surname or role, never "Interviewee" in one place and a name in another.
Anonymising a transcript
If you promised anonymity, the transcript is where you deliver it. Replace names, employers, place names and any detail that identifies the person in a small field, job titles in a niche industry are often more identifying than the name.
Keep the unredacted version separate and secure, and note in the header which version this is. Researchers get caught out by circulating the wrong file more often than by bad redaction.
From recording to formatted transcript
Upload the interview and voxnoto returns it split by speaker with timestamps, which is the skeleton of the template above. Rename each speaker once and the label applies throughout.
Then do the part a machine shouldn't: mark what's genuinely inaudible rather than accepting a plausible guess, and check names and technical terms against the audio.
Questions and answers
How should an interview transcript be formatted?
A header with the interview details, then timestamped turns with consistent speaker labels.
What does clean verbatim mean?
Fillers and false starts removed, wording unchanged.
How do I mark something I can't hear?
[inaudible] with the timestamp. Never guess silently.
Should transcripts include timestamps?
Yes, at least at each speaker turn, so quotes can be checked against the audio.
How do I anonymise a transcript?
Replace names and identifying details consistently, and keep the original separately.
Can I get a formatted transcript automatically?
Yes. Upload the recording and edit it into this template.
Did not find your answer?Browse all questionsEmail support
