More reliable imports for phone audio and irregular timestamps
Recordings with tricky duration data or slightly out-of-order word timing now complete more reliably.
- Reliability
- Audio
- Timestamps
What's new
New ways to work with recordings, thoughtful improvements, and fixes you can feel.
Recordings with tricky duration data or slightly out-of-order word timing now complete more reliably.
A transcript now becomes available as soon as the words are ready, while AI notes finish separately.
Pepys now accepts public Rumble, Twitch, and Kick links alongside the platforms already supported.
Answers include supporting transcript moments so you can check the wording and jump back to the source.
Choose a data region, keep a recording when you need it, or remove source media while keeping the transcript.
Give Pepys important names and terms, then export timed files for editing or qualitative research.
Set a recording's language before it starts, or correct the language from the finished transcript.
Find, sort, edit, and manage transcripts from a faster library that works properly on mobile.
Start a transcription from the page you are viewing, without copying the link into another tab.
Remove filler words and repeated starts while keeping the cleaned text tied to the recording.
Send recordings from your own workflow, receive completion updates, and export finished transcripts by API.
Create a public transcript link, embed it elsewhere, or run speaker detection again when a recording needs a closer pass.
Search inside a transcript, jump back to the exact moment, and ask follow-up questions without rereading everything.
Export transcripts as documents, subtitles, or structured data, with speaker names kept consistent.
Upload a recording or paste a link, then get a timestamped transcript with speaker labels and useful notes.