Podcast Audio Cleaner
Level the loudness and cut the long silences and filler words from a podcast recording, and get the cleaned audio and the list of every cut. Upload an MP3, WAV or M4A of one episode (up to 60 minutes and 100 MB).
What installing it brings
- 1 workflow39 steps in what runs behind it.
- 1 screenThe pages of its application.
- 1 tableThe data tables it works with.
About this application
Level the loudness and cut the long silences and filler words from a podcast recording, and get the cleaned audio and the list of every cut. Upload an MP3, WAV or M4A of one episode (up to 60 minutes and 100 MB). The app gets a transcript with word timestamps from AssemblyAI, cuts the pauses longer than your threshold (leaving a natural breath), cuts the hesitations of your list (um, uh, euh and your own words), trims the dead air at the start and the end, leaves long gaps such as a music bed or an ad alone, and brings the whole episode to your target loudness (-16 LUFS stereo or -19 mono by default). You get the clean MP3, how much time was cut, and every cut with its time and reason, as a list you can copy or download as CSV. Each episode becomes a card on your Episodes board (Needs a listen, To review, Approved, Published); the same file cleaned again updates its card. Settings: language, target loudness, shortest pause to cut, pause left after a cut, words to always cut, gaps over 12 seconds, cut list only. Needs an AssemblyAI account connected with your own API key. The app does not remove background noise, the loudness is a target and not a measurement, one pass makes at most 48 pieces (the longest cuts first; run the cleaned file again to cut the rest), and nothing is published anywhere.
The published app, exactly as it runs.
This preview is not interactive.
The workflow
What runs behind it: 39 steps, 1 screen.
Published by
livecontext
@livecontext