Audio to text with the speakers separated.
Drop a recording — MP3, WAV, M4A, FLAC, OGG, or a link to one — and get back text you can actually use: punctuated, paragraphed, labelled by speaker, and timed to the word.
What you get
What separates a transcript from a text dump.
Punctuation and paragraphs
Raw speech recognition returns an unbroken run of lowercase words. Every transcript here comes back with sentence breaks, capitalisation, and paragraphing, so it reads like written English instead of a stenography exercise.
Who said what
Diarization runs on every job. Two-person interviews, three-way panels, and noisy roundtables come back split into labelled turns rather than merged into one voice.
Timed to the word
Every word carries its own timestamp, not just the segment. That's what makes click-to-seek work, and what makes SRT and VTT exports line up properly instead of drifting a second late.
Any format you have
MP3, WAV, M4A, AAC, FLAC, OGG, WMA, AMR and more — plus the audio track of any video file. No converting to a supported format first, and no 25 MB ceiling.
Why not just use a free web converter
Free converter vs real transcript.
✗ A typical free online converter
Fine for a short voice note. The limits show up the moment the recording has two people in it or runs past ten minutes.
- Short file caps, then a paywall
- One speaker assumed, always
- No word-level timing
- TXT only, if you're lucky
- Audio often retained and reused
✓ A Whipscribe transcript
Built for recordings that matter: interviews, calls, lectures, field tape.
- Up to 5 GB and 12 hours per file
- Speaker turns labelled automatically
- Word-level timestamps throughout
- TXT, SRT, VTT, DOCX, JSON
- Never used to train AI
Sample output
Speaker-labelled. Click-to-seek. Exportable.
An interview recording, as it comes back — labelled, punctuated, and timed.
Export
One transcript. Five clean formats.
Every pack exports all five — no format is held back for a higher tier.
Plain text
De-ummed paragraphs. Ready to paste.
SRT captions
Word-level. Every video editor reads this.
WebVTT
HTML5 player + YouTube uploads.
Show notes
Formatted with chapters and pull-quotes.
Machine-readable
Per-word timing + speaker IDs.
Pricing
Honest pricing, no surprises.
No subscription and no auto-renew — every tier is a one-time minute pack, and credits never expire. Buy once, use the minutes whenever you need them.
Casual pack
$4/300 min
5 hours of audio. One-time — credits never expire.
- 300 minutes of audio
- All export formats
- Unlimited Claude chat per transcript
- Executive summary & key quote
Best value
$12/2,000 min
33 hours of audio — $0.36 an hour. Credits never expire.
- 2,000 minutes of audio
- Priority queue
- All export formats
- Unlimited Claude chat per transcript
Bulk pack
$24/5,000 min
83 hours of audio — $0.29 an hour. Credits never expire.
- 5,000 minutes of audio
- Priority queue
- API access for batch pipelines
- All export formats
FAQ
Audio-to-text questions, answered.
Which audio formats work?
MP3, WAV, M4A, AAC, FLAC, OGG, WMA, AMR, and most other common formats, plus the audio track of any video file. If it plays, it almost certainly converts.
How accurate is it?
Whisper-class: typically under 5% word error rate on clean English speech. Accuracy drops on heavy background noise, overlapping speakers, or thick accents — as it does for every engine, human transcribers included. Fixing two words by hand beats fighting a worse model.
How long does it take?
Most one-hour recordings come back in a few minutes. The job runs server-side, so you can close the tab and come back to it — nothing is happening in your browser.
Can it tell speakers apart?
Yes, on every job. Diarization labels each turn, and you can rename SPEAKER 1 to a real name once and have it apply throughout the transcript and every export.
Is my audio kept?
It is never used to train any model. We run Whisper on our own GPUs precisely so audio doesn't get handed to a third-party API, and you can hard-delete any file and its transcript from your account.
Related
Related tools and pages.
Drop a recording. Read it in minutes.
Try WhipscribeOperated by Neugence Technology Pvt. Ltd. · contact@neugence.ai · Security · Privacy · Terms