Journalists
Quote-check a 40-minute call in the time it takes to skim the summary. Timestamps make every line defensible.
Transcription
Every word.
Every word.
Every speaker.
On the record.
Interviews, meetings and lectures turned into structured transcripts — speaker labels, timestamps, and a search that lands on the exact sentence.
30 free minutes · Speaker labels included · Export TXT · SRT · VTT
interview_01.m4a
04:12
A
So the thing we underestimated was how long the migration would really take.
04:26
B
Right — and that pushed the roadmap back by a full quarter, not two weeks.
04:38
A
Which is exactly why I want the roadmap reviewed every Friday now.
04:51
B
Agreed. I will own that cadence and send the notes after each session.
05:03
A
Perfect. Then let us talk about the hiring plan.
Speaker A · 54%
Speaker B · 46%
Export ▾
94%+
word accuracy on clear audio
~1 min
for a 1-hour recording
A / B / C
speaker turns, labelled
4 formats
TXT · SRT · VTT · ASS
How it works
From raw audio to a file someone can quote.
No plugin, no desktop install. The whole loop happens in one tab, and nothing is charged until the transcript exists.
-
01
Upload the recording
Interviews, calls, lectures, field notes — MP3, M4A, WAV, MP4 and more, up to 2 hours. The video stays local; only the audio travels.
-
02
Voices get separated
Before a single word is written, the audio is split into speakers. Turns are labelled A, B, C, so a two-hour panel does not read as one long paragraph.
-
03
Review inside the timeline
Click any sentence to jump to that second of audio. Fix a name, merge a false start, and the timestamps reflow with you.
-
04
Hand it off
Export TXT for notes, SRT or VTT for captions, or send the transcript straight into AI Translate when the audience is not English-speaking.
“We stopped treating transcripts as a by-product. Now the transcript is the record, and the recording is the backup.”
— Research lead, a 30-person user-lab team
Who it is for
Built for anyone who has to prove what was said.
Researchers
Code interviews without re-listening. Speaker labels keep each participant distinct through the whole session.
Podcasters
Transcripts double as show notes, SEO pages and searchable archives — produced from the same file you already published.
Support & sales
Turn call recordings into reviewable records. Search a customer name and land on the exact sentence.
Questions
Frequently Asked
How does speaker separation work?
Voices are clustered before transcription starts, then each sentence is attached to the speaker who produced it. Clear, unmixed audio works best; two people on one microphone will still be separated, but expect the occasional swap.
Are timestamps accurate enough to cite?
Timestamps are derived from word-level alignment, so they land within a fraction of a second. For formal citations we still recommend a quick listen at the marked moment.
What happens to my audio after the job?
Only the extracted audio is sent for processing. It is released once the transcript is returned, and the original video never leaves your browser.
Can I search inside a long recording?
Yes. Search runs across the whole transcript and highlights every hit, with one click jumping the player to that second.
Put the recording down. Pick the transcript up.
30 free minutes to test with a real file — the messy, crosstalk-heavy one.