How to transcribe a set of long research interviews and build a quote bank you can play
You finished fieldwork. On your laptop sit fourteen interviews, each between one and three hours, as M4A and WAV files from a recorder and a few MP4s from video calls. Somewhere in those twenty-odd hours are the forty quotes your chapter will rest on. This guide is about getting from that folder to a quote bank where every excerpt carries the second it was spoken, so you can always go back and hear it in context before it goes into a paper.
Have a recording like this? Packs from $4, credits never expire → timestamps, overview. Summary free to read.
The short version
- Name the files before you upload so each one carries the participant code and date, for example
P07_2026-08-14.m4a. - Upload each interview as one file. Up to 10 hours or 5 GB per file, so a three-hour session never needs splitting. A one-hour file is usually done in about two minutes.
- Skim the Intelligence tab first: the summary, chapters with timestamps and key quotes give you a map of each interview before you read a line of it.
- Code against timestamps, not page numbers. Ask AI Chat where a theme appears, click each timestamp, listen, and only then copy the quote.
- Export the transcript as DOCX or TXT for your analysis software, and ask AI Chat for an Excel quote table that lands in the recording's Resources.
Why long interviews need a different workflow
A 40-minute interview can be read start to finish in one sitting. A 2½-hour life-history interview cannot, and a set of fourteen of them is a book-length corpus. Three things go wrong when researchers treat long interviews like short ones.
- Quotes lose their context. A line copied into a codebook without a timestamp is impossible to recheck. Six months later, a reviewer asks whether the participant was being ironic, and you cannot find the moment.
- Themes drift across the set. What you coded as "distrust of institutions" in interview 2 becomes "bureaucratic fatigue" by interview 11, and you only notice when writing up.
- Hearing beats reading for tone. Hesitation, laughter and a long pause before an answer are part of qualitative data. A transcript you can click to hear keeps that within reach without relistening to whole sessions.
Step by step: from audio files to a coded quote bank
1. Prepare the files
Keep the original recordings. If your recorder made a WAV, upload the WAV rather than a compressed copy someone sent over chat. Give each file the participant code and date so your library sorts itself, and keep the key that links codes to real names offline, in the place your ethics protocol specifies.
2. Upload one interview at a time, on desktop
Open whipscribe.com on your computer, drop the file onto the page and enter your email when asked after the upload. Accepted formats are .mp3, .wav, .mp4, .m4a, .flac, .ogg, .webm and .mov. Large files are fine; a three-hour WAV is a normal upload. The language is detected automatically, and 99+ languages are supported, which helps when a participant switches into their first language for a passage.
3. Read the map before the territory
Each transcript gets an Intelligence tab with a summary, chapters with timestamps and key quotes. On a three-hour interview, the chapters are the fastest way to see where the participant talked about childhood, work, the turning point and the present. Use them to decide which stretches to read closely first. They are a starting point for your own reading, not a substitute for it.
4. Code with the audio one click away
Every line of the transcript has a timestamp you can click to play the audio from that second. Press ⌘K (or Ctrl K) to find a word, and every match is highlighted. When you are looking for an idea rather than a word, ask AI Chat: "Where does she describe the first time she asked for help?" The answer quotes the lines it found with timestamps you can click to hear them.
5. Build the quote bank
Drag across the lines you want and copy them with their timestamps into your codebook. For a first pass across one interview, ask AI Chat to "make an Excel table of every passage about trust, with the timestamp and a one-line description". It saves the file to the recording's Resources, and you edit it from there. The table is a draft for you to check, not a finished analysis.
6. Export for your analysis software
Download the full transcript as DOCX or TXT with timestamps and import it into the qualitative analysis tool or word processor you already use. Keep the participant code in the file name so every coded segment can be traced back to one recording and one second.
Start with your longest interview Upload the audio file and get a transcript you can click to play Packs from $4 for 500 minutes. Credits never expire. Upload an interview →A worked example
A doctoral researcher with 14 interviews on returning to work after illness
Fourteen semi-structured interviews averaging 1 hour 40 minutes, about 1,400 minutes in total, recorded on a handheld recorder and two video calls. She bought the $12 pack (2,000 minutes) and uploaded one file per evening. For each, she skimmed the chapters, then asked AI Chat "where does the participant talk about their manager?" and "where do they describe the first week back?". In P09, a 2 h 51 min interview, the answer pointed to 1:47:12, a passage she had forgotten, where the participant described rehearsing the conversation with her manager in the car park. She clicked, listened, and copied the four lines with their timestamp into the "anticipation" code. Across the set she built a bank of 112 timestamped excerpts. The remaining 600 minutes stayed on her account for follow-up interviews.
The example is invented, but the shape is typical: the slow part of qualitative work is finding and rechecking, not typing.
Recording tips that pay off at analysis time
What you get for each interview
- A full transcript where every line has a timestamp you can click to play the audio from that second.
- An Intelligence tab with a summary, chapters with timestamps and key quotes.
- AI Chat that answers questions about the interview with timestamps you can play, and can make Word, PDF, Excel or SRT files that are saved to Resources.
- Downloads in TXT, SRT, VTT and DOCX.
Price: $4 buys 500 minutes, enough for about four two-hour interviews. Larger sets: $8 for 1,000 minutes, $12 for 2,000 minutes, $24 for 5,000 minutes. Credits never expire and you pay by card through Paddle. See how to summarize 20 interviews for working across a whole set, and the academia use case for how other researchers use it.
Frequently asked
How long can one interview recording be?
Up to 10 hours or 5 GB per file, so a three-hour interview goes up as one file. You do not need to split it, and the timestamps run continuously from the start of the recording to the end.
What does a set of 15 two-hour interviews cost?
Fifteen two-hour interviews are 1,800 minutes. The $12 pack covers 2,000 minutes, so the whole set fits in one pack with 200 minutes left over. Credits never expire, so the remainder waits for your follow-up interviews. Smaller sets can start with the $4 pack for 500 minutes.
Which file types can I upload?
MP3, WAV, MP4, M4A, FLAC, OGG, WEBM and MOV are all accepted, which covers phone voice memos, dedicated recorders and video calls saved to disk. Upload the original file rather than a re-compressed copy when you have it.
Can I get the transcript into my analysis software?
Download the transcript as DOCX or TXT with timestamps, which most qualitative analysis tools and word processors can import. SRT and VTT are also available if you want to keep the timing in a subtitle format. In AI Chat you can also ask for a Word, PDF or Excel file, such as a table of quotes with their timestamps, and it is saved to the recording's Resources.
Can the AI code my interviews for me?
It can help you get there faster, but the coding stays your judgement. Ask AI Chat where a theme comes up and it answers with the lines it found and timestamps you can click, so you check every quote against the audio before it goes into your codebook.
Is my interview audio used to train AI models?
No. Recordings are transcribed on our own servers and are never used for training. You can delete a recording and its transcript at any time, which matters when your ethics approval sets rules on how long audio is kept.
Your interviews are already recorded. Upload the first file, read the chapters, and start a quote bank where every excerpt can be heard again in one click. $4 for 500 minutes, credits never expire.
Upload your interview file →