Smart Audio Prep ("Clean up audio")

Overview#

A drawer on the Ready to transcribe card (opened with Edit) that shows the recording as a waveform and lets you remove long silences or marked stretches before transcription. Everything happens in the browser.

Purpose#

To avoid paying for silence (1 credit = 1 minute) and to cut irrelevant parts before upload.

Who uses it#

Anyone preparing a recording, most usefully for long recordings with breaks.

Screenshot#

Interface explanation#

• Original / Result: switch between the recording as uploaded and the cleaned version.
• Waveform with a time ruler; Play, Zoom out (−) / Zoom in (+).
• Mark start / Mark end: mark a stretch to cut by hand.
• Summary row: Original (2:49), Removing (0:00), Final (2:49), Saves (0 credits).
• Status line: "No long silences found." for the test file.
• Download (the cleaned file) and Use cleaned audio; both are disabled until something is removed.

See it in action#

A recording with a long silence in the middle, cleaned on the device, recorded on the real app.

How to use#

1. On Ready to transcribe, press Edit (scissors).
2. Review the suggested cuts, or mark a stretch with Mark start and Mark end.
3. Check Saves N credits, then Use cleaned audio.
4. Back on the card, a Cleaned note appears with Restore to undo (VERIFIED THROUGH CODE).

Example scenario#

The synthetic meeting had no long pauses, so the drawer reported "No long silences found." and nothing needed removing.

Expected result#

When cuts are applied, the card's length and credit price drop to the Final duration.

Important notes#

• Drawer VERIFIED THROUGH UI. Applying cuts, the "{mm:ss} of silence found → Clean up" hint and Restore are NOT VERIFIED in the UI: the test file had no silence long enough to trigger them.
• Per the source, recordings over an hour can be edited only on devices with enough memory ("Recordings over an hour can be edited on a computer.").