Podcasts, voice notes, voiceover
If the file is audio-only, there is no picture to protect — but the same privacy rule applies: the file is decoded in the browser, a 16 kHz mono track is sent for word timings, and the download is encoded on your machine. That is enough for a podcast episode under the comfortable 10-minute / 100 MB ceiling. Longer shows need to be split first; this is a browser tab, not a studio DAW.
Fillers and silence in one pass
People searching “remove filler words from audio” and people searching “remove silence from audio” are often cleaning the same take. Here both run together. Lower the silence slider if you want tight jump-cuts. Raise it if you want room tone between sentences. Ums are independent of that slider: they are cut because they are filler tokens, even when they sit inside otherwise continuous speech.
What you get back
Preview never re-encodes; it just skips ranges. When you like the edit, Save writes M4A. There are no stems, no loudness-matching, and no “enhance” model. If you need Adobe Podcast Enhance, that is a different product. This page only removes the words and pauses you do not want.