Transcribe research interviews to text
Upload your interview recording. Get a text transcript. Free for interviews up to 30 minutes: 1 a day without an account, 3 with a free account.
Files deleted after processing · 99 languages · First file without sign-up
Drop your file here
or click to browse files
First 30 min free · full file US$5, one-time
MP4, MOV, MKV, WebM, MP3, M4A and more · up to 2 GB
The time math
A peer-reviewed study (NIH, 2024) confirmed manual transcription takes 3–10 hours per interview hour.
3–10 hours saved per interview
Manual transcription of a 60-minute interview takes hours. In October 2026, half of the files on our servers were transcribed in 14 seconds or less, and a 145-minute recording took 2 minutes 45 seconds. Times don't include uploading, which depends on your connection. You spend the time reviewing, not typing.
Deleted after processing
Your file is deleted from our storage after processing, and we don't store transcripts. If you take the free 30-minute preview of a longer file, we keep that upload for up to 24 hours so you can finish it without uploading again. Check your IRB protocol: audio is processed by our transcription providers (see the privacy policy).
Accurate on clear recordings
Powered by Whisper large-v3 turbo. Clear interviews with one or two speakers transcribe well; budget review time before quoting.
Export and edit in Word
Download the transcript as TXT and open it in Word or Google Docs. Add speaker labels, clean up errors, and format for your NVivo, MAXQDA, or Atlas.ti import.
Questions from researchers
How long does it take to transcribe a research interview?
In October 2026, half of the files on our servers were transcribed in 14 seconds or less, and a 145-minute recording took 2 minutes 45 seconds. Times don't include uploading, which depends on your connection. Files longer than 30 minutes need Pro or a Week Pass.
Is my interview data kept private?
Yes. Uploaded audio is sent only to our transcription providers (Groq, with Cloudflare Workers AI as a fallback) and deleted from our storage after processing (a free-preview upload is kept for up to 24 hours). We don't sell or share recordings for any other purpose or use them for model training, and we don't store transcripts. The transcript is shown only in your browser.
How accurate is automated transcription for research interviews?
TranscribTxt uses Whisper large-v3 turbo, which is accurate on clear one-on-one recordings. Accuracy drops on multi-speaker recordings with overlapping speech, recordings with heavy background noise, or strong regional accents. Budget 20-30 minutes of review time per hour of audio to correct errors before using the transcript in research.
Can I cite a machine-generated transcript in academic research?
You can use the transcript as a working document and cite your primary source (the original interview). Most institutions require researchers to review and verify automated transcripts before treating them as research data. TranscribTxt gives you an accurate draft — you review it for errors.
What file formats work for research audio?
MP3, WAV, M4A, FLAC, and OGG all work. WAV is the most common format from digital voice recorders used in field research. MP3 is fine for most purposes. For sensitive interviews, WAV at 48kHz gives the best transcription accuracy.