Research interview transcription, ready to code and cite
Upload the recording or paste a link – get a speaker-labeled transcript plus the themes, verbatim quotes, and Q&A breakdown that turn raw audio into analysis.
60 min free · no card required · we never train on your audio
How do you transcribe interviews?
To transcribe a research interview, upload the audio or paste its link and Pepys returns a speaker-labeled, timestamped transcript in minutes – plus AI-surfaced themes, verbatim quotes, and a question-by-question breakdown you can drop straight into your coding and write-up. It's pay-as-you-go with no subscription, and credits never expire.
Made for researchers
Every interview you run is data trapped in audio. Before you can code a single theme, pull a quote you can stand behind, or write up a finding, someone has to turn hours of recording into accurate, attributed text – and that transcription step is the slowest, least fun part of qualitative work. Pepys does it the moment you upload, with speakers separated and timestamps you can cite, so you spend your time on analysis instead of typing.
The grind comes later, in the coding. You build a codebook, tag interviewer and participant turns, and chase one recurring idea across a dozen sessions before you can claim it's a pattern. Research interview transcription with diarization and word-level timestamps means each verbatim line carries the exact second it was said, so a quote drops into your write-up already attributable and the in-transcript search finds every mention of a concept without you scrubbing the audio. Export to JSON for NVivo or Atlas.ti, or DOCX for the appendix.
Coding-ready transcripts
A clean, speaker-separated transcript you can import straight into NVivo, Atlas.ti, or a spreadsheet to start tagging themes.
Verbatim, attributable quotes
Every line is timestamped, so the quote you cite in a paper links back to the exact moment a participant said it.
Searchable across a study
Find every mention of a concept across dozens of sessions instead of scrubbing recordings one by one.
Multilingual fieldwork
Detects the spoken language automatically across 99+ languages, so cross-country studies don't need a separate vendor per locale.
Built in, not bolted on
Recurring themes, supporting quotes, and a participant recap
Every interviewis analyzed automatically the moment it’s transcribed. Here’s a real sample, run through it.
Plain Words, Please: An Older Adult on Why a Banking App Made a Simple Transfer Feel Risky
In a qualitative usability interview, an older participant walks through moving money between two accounts in a banking app and describes friction at every step: a dense home screen with no visible transfer label, an icon they couldn't decode, account names shown as codes instead of the nicknames they had set, and a green flash with no readable receipt. A ten-second task took four minutes and left them checking both balances by hand. Their core ask is plain language, recognizable account names, and a confirmation screen that stays put – a usability gap, not a technology gap.
Themes
Notable quotes
- “I'm looking for the word transfer and it isn't there.”
- “I'm thinking, am I about to press the wrong thing and send money to a stranger? That fear is always there with money.”
- “I guessed. And I hate guessing with money.”
- “It took me maybe four minutes to do a thing that should take ten seconds.”
- “It just flashed something green for half a second and went back to the home screen. No receipt, no nothing.”
- “I'm not slow with technology. The app is just talking in a language that wasn't built for me.”
Q&A
What was the participant's first reaction when opening the app to make a transfer?
They froze for a second, faced with a dense screen of numbers and icons, and could not find the word transfer – it was hidden behind an icon they described as two arrows doing nothing in particular.
Why did selecting the accounts cause friction?
The from and to accounts were shown as codes rather than the nicknames the participant had set, so they had to guess and selected the account by the last four digits, the only part they recognized – stretching a ten-second task to about four minutes.
How did the participant feel about the transfer confirmation?
They felt no confidence at all: the app flashed green for half a second and returned to the home screen with no readable receipt, so they checked both balances by hand to be sure the transfer had gone through.
What single change did the participant ask for?
Plain language and recognition over recall – use the word transfer, show the account nicknames they set, and provide a confirmation screen that stays put until they close it. They framed the problem as the app's, not their own.
Clean, speaker-labeled, click-to-seek
Ask, don’t scrub
Ask the transcript anything.
An hour-long recording? Don’t skim it – ask. Every answer stays grounded in your transcript and cites the exact timestamp, so you can jump to the moment and check it yourself.
What barriers did the participant hit when trying to make a transfer?
On opening the app they froze, facing a dense screen of numbers and icons with no visible 'transfer' word – it sat behind what they called two arrows doing nothing in particular. Then the from and to accounts showed as codes instead of the nicknames they'd set, so they guessed and picked the one ending in the four digits they recognized.
How did they feel after sending it, and what change did they ask for?
They felt no confidence – the app flashed green for half a second and returned home with no receipt, so they checked both balances by hand. Their one ask was plain words: say 'transfer', show the account nicknames, and give a confirmation screen that stays put until they close it.
Grounded in your transcript – if the answer isn’t in the audio, it says so instead of guessing.
Who said what
Speaker labels that survive cross-talk
Automatic speaker diarization. Two people, four people, cross-talk and interruptions – interviews, panels, messy meetings. Pepys keeps each voice on its own line instead of blurring them into one, so you never rewind to figure out who was talking.
So the festival nearly didn't happen this year–
–it almost didn't. We lost the venue three weeks out.
Three weeks? How do you even start to–
You call everyone you know. The whole town pitched in.
And that's how it ended up in the park.
Record in any language – 99+ detected automatically
- English
- 中文
- Español
- العربية
- हिन्दी
- Français
- 日本語
- Português
- Русский
- Deutsch
- 한국어
- Italiano
- বাংলা
- Türkçe
- فارسی
- Tiếng Việt
- தமிழ்
- Polski
- ไทย
- Українська
- Nederlands
- עברית
- Ελληνικά
- తెలుగు
- Bahasa Indonesia
- اردو
- Svenska
- मराठी
- Română
- Magyar
- Čeština
- ગુજરાતી
- Kiswahili
- ქართული
- Tagalog
- አማርኛ
Works with the platforms you live in.
Paste a link from YouTube, TikTok, Instagram, Facebook, Spotify, or Apple Podcasts – or drop in any audio or video file. We transcribe it once, then you export it however your workflow needs.
- YouTube
- TikTok
- Spotify
- Apple Podcasts
- or any file
Export to any format
- TXT
- Markdown
- DOCX
- SRT
- VTT
- JSON
Most useful for researchers: TXT · DOCX · PDF · SRT · VTT · JSON
Timestamps, speaker labels, and subtitle timing carry through to every export.
How research interview transcription works
Upload or paste a link
Drop your interview or paste its link – any audio or video, in any language.
Get your transcript
A clean, speaker-labeled transcript with AI notes tuned to your format, ready in minutes.
Edit and export
Fix anything inline, then export to SRT, VTT, TXT, DOCX, PDF, or JSON.
Why researchers pick Pepys
No subscription – pay per interview, and the credits never expire, so a study that spans a whole grant year works fine.
We never train on your audio, which matters when participants are identifiable and your IRB protocol says so.
Speaker labels and timestamps come standard, so quotes stay attributable and citable.
Themes, quotes, and a Q&A breakdown are drafted from the transcript, not a separate paste into another tool.
What researchers say
I work across three languages and it detected each one correctly without me changing a single setting. The timestamps line up to the word – exactly what my research needs.
Priya N.Linguistics PhD candidate · Trustpilot Every user interview comes back as a clean, searchable transcript I can tag and quote directly in my reports. Synthesis used to be the slowest part of my week and now it's an afternoon. The speaker labels alone are worth it for me.
Sofia L.UX researcher · G2 multilingual focus groups, transcribed and translated into one working language so i can compare responses side by side. used to wait days on a vendor for this – now its same-day.
Erica B.Market researcher · Product Hunt
Research interview transcription – questions, answered
How do I transcribe a research interview?
Upload the recording or paste its link and Pepys returns a speaker-labeled, timestamped transcript in minutes, along with AI-surfaced themes, verbatim quotes, and a question-by-question breakdown you can take straight into coding.
Can it separate the interviewer from the participant?
Yes. Speaker diarization labels each voice, so a one-on-one interview comes back as clear interviewer and participant turns rather than one undivided block. You can rename a speaker once and it updates throughout the transcript.
Is the transcript accurate enough to quote in a paper?
It is verbatim and timestamped, so any line you cite is tied to the exact moment it was said. Anything the model gets wrong with a name, an accent, or a technical term you can correct inline before you export.
Do you use my interview audio to train AI models?
No. We never train on your audio. For human-subjects work where participants are identifiable and your ethics approval restricts data use, that point is the reason a lot of researchers choose Pepys.
Can it handle interviews in other languages?
Yes. It auto-detects the spoken language across 99+ languages, so multilingual or cross-country fieldwork comes back transcribed without you changing a setting for each session.
What can I export to start analysis?
Plain text, DOCX, and PDF for write-ups, JSON for importing into qualitative software, and SRT or VTT if you need timed captions for an archived recording. One click each.
Do I have to subscribe?
No. Pepys is pay-as-you-go – buy a block of minutes, use them across a study whenever the work happens, and the credits never expire. You can start free with 60 minutes, no card.
More industries
Turn your next interview into a coded-ready transcript, themes, and citable quotes – and pay only for that interview.
Pay as you go – credits never expire, nothing to cancel. Or start free with 60 minutes, no card.