🎧
Audio and video to text
Meetings, interviews, lectures, podcasts, videos: a transcript with times and speakers, ready to edit, to subtitle, or to turn into minutes.
Billed by use: an hour of recording is usually a few credits.
A real run
What was entered
- Recording (AI narration made on this site)
What came back, unedited
00:00 Most people think good coffee starts with expensive beans. 00:05 It starts with water. 00:06 A cup is about ninetyeight percent water. 00:09 Hard tap water full of minerals flattens the taste; 00:13 very soft water makes it sour. 00:16 Use filtered water heated to around ninetythree degrees Celsius. 00:21 Grind the beans just before brewing. 00:24 Those two changes cost almost nothing. 00:28 They make more difference than a new machine.
Run on 2026-10-18 through this page’s own request, once per language; nothing edited.
How it works
- 1Upload a recording or a video; long files are fine.
- 2Wait while it is read: a minute or two per half hour of sound.
- 3Fix any word in place, name the speakers, then export or keep working on it.
Questions
- Which languages?
- Chinese (including common dialects), English, Japanese, Korean and the major European languages are read well; the transcript stays in the language spoken.
- How accurate are the times?
- To about a second, per sentence. Good for finding a passage and for subtitles; check the subtitles against the video before publishing.
- Is my recording kept?
- The file is stored in your account so it can be read, and the transcript is saved to your library; neither is published or shared.
- Can it make minutes or a summary?
- Open the transcript in the document workbench and ask for minutes, a summary or action items; every change is shown before it lands.