Find the interview in the recording
An hour of conversation usually holds twenty good minutes. Paraphrase transcribes all of it with a timestamp on every word, so you can read for the answers worth keeping and cut the rest straight from the text.
Why Paraphrase?
The words are already a map of the recording. Editing them is faster than finding each one in a waveform.
Built for conversations
For the twenty good minutes inside an hour of talking.
Upload the recording
Bring in the file from your recorder or your call, in most audio and video formats.
Several sessions, one project
Add each session as a take and edit them as one transcript.
Clean up with AI
One button reads the transcript and proposes the cut: filled pauses, false starts and the sentence you said twice. It shows you what it took, and you can put back any word it removed.
Undo and redo
Ctrl+Z steps back through your edits and Ctrl+Shift+Z steps forward again.
Export the way you publish
WAV or FLAC keep every sample of a lossless master. MP3, M4A and Opus are there for publishing.
Never used for training
Your recordings and transcripts are never used to train any model, ours or anybody else's.