Interview transcript template
Download a free interview transcript template as a Word file or plain text. It covers what a usable interview transcript needs: the recording details and consent, a key for speaker labels, marks for words you can't hear, a timestamp on every speaker turn, and a log for checking quotes against the audio before you publish.
- Word (.docx) and plain text (.txt)
- Speaker key and notation marks
- Quote check with timestamps
Download the template
Free, and no account needed. Both files have the same sections.
The Word file opens in Microsoft Word, Google Docs and LibreOffice.
What's in the template
Interview details
Who, when and where, the recording file and its length, and how consent to record was given.
Speaker key
Short labels such as INT and R1, matched to names and roles. Codes keep a research transcript anonymous.
Notation
Marks for inaudible words, crosstalk, pauses, laughter, [sic] and redactions, so everyone on a project writes them the same way.
Transcript table
One row per speaker turn: the time, the speaker, and what they said.
Quote check
Each quote you plan to use, its timestamp, and whether you listened to it again.
Follow-ups
Facts to confirm, spellings to check and questions for the next interview.
Clean verbatim or full verbatim?
Decide before you start and write the choice in the template. A transcript that mixes the two makes every quote harder to trust.
| What happens to | Clean verbatim | Full verbatim |
|---|---|---|
| Filler words (um, uh, you know) | Removed | Kept |
| False starts and repeated words | Removed | Kept |
| Pauses, laughter, other sounds | Left out unless they change the meaning | Marked, for example [pause] or [laughs] |
| Grammar slips | Tidied, with the meaning unchanged | Left as spoken, with [sic] where it helps |
| Usual uses | Journalism, user research, business interviews | Legal records, linguistics, research where how something was said matters |
AI transcripts, GhostMinutes included, often leave out filler words and false starts, so they read closer to clean verbatim. For full verbatim, listen through and add what is missing.
Speaker labels that hold up
Pick one label scheme and keep it for the whole project. INT for the interviewer and R1, R2 for respondents works well in research. Journalists often use first names.
If someone takes part anonymously, keep their real name out of the transcript, details block included. Store the key that links codes to people in a separate file with tighter access.
GhostMinutes labels voices Speaker 1, Speaker 2 and so on. Rename a speaker once and every line changes, exports included. If a line lands under the wrong person, reassign it.
Timestamps for checking quotes
Put a timestamp on every speaker turn. When an editor, or the person you interviewed, questions a quote, you can jump straight to it instead of scrubbing through the whole recording.
Use the time from the recording file, not the clock time of the interview. That way the stamps still match if the recording started late.
Before you publish, copy each quote into the quote check, play the audio at that time, and mark it checked. Names, numbers and technical terms are where transcripts go wrong most often, whether a person or a model typed them.
Fill it in with GhostMinutes
- 1
Upload the recording
Use the box on interview transcription or the upload page. Files can be up to 1 GB.
- 2
Name the speakers
Rename Speaker 1 and Speaker 2 to match the labels in your speaker key.
- 3
Export with timestamps
Export DOCX or TXT with timestamps and speaker names left on. Each line starts with the time, like [4:07], and the speaker.
- 4
Paste it into the template
Paste the lines into the transcript section, fill in the details, then work through the quote check.
Interview transcript questions
Transcribe the interview first
Upload the recording, rename the speakers and export a DOCX with timestamps.
Upload an interview