Filler word remover · Mac and iPhone
Cut the ums, uhs and likes out of your video
Filler words are the habit that makes a confident speaker sound unsure, and cutting them by ear takes longer than the recording did.
Zella turns your speech into a word-level transcript on the device, spots the fillers in it, lists the cuts it plans to make and removes them with the gaps closed. Nothing is uploaded, and every cut can be undone.
- Mac and iPhone
- Free on every tier
Quick answer
Zella removes filler words on your device: Remove Fillers on the Mac and Clean Up on iPhone find um, uh, like, you know and sentence-start fillers in the transcript, cut them and close the gaps. On iPhone a verbatim on-device Whisper model writes every um and uh out, so they really get cut; on the Mac, Apple's speech engine often leaves plain um and uh out of the transcript, so soft and sentence-start fillers are what it catches reliably. It is free on both.
How to do it in Zella
-
Step 1: Open the take in the editor
Record or import a clip with speech. If you want both, run silence removal first: silences, then fillers, is the order Zella recommends.
-
Step 2: Mac: click Remove Fillers
In the sidebar's Auto Tools tab, click Remove Fillers. Zella transcribes, spots likely fillers and lists each cut, such as like, or so at the start of a sentence.
-
Step 3: iPhone: tap Clean Up
Clean Up removes filler words and silences together, working from the verbatim on-device transcript.
-
Step 4: Review, then apply
Check the list before you apply. Start light, then run it again if you want a tighter cut.
-
Step 5: Trim any stragglers
On the Mac, split out any plain um the speech engine left out of the transcript by hand on the timeline. On iPhone, delete the word in the transcript and the video ripple-cuts with it.
Why do it this way
Three kinds of filler
Hard fillers (um, uh), soft fillers (like, you know) and sentence starters (so, OK, I mean) are all on the list.
Honest about each platform
iPhone transcribes verbatim, so plain um and uh are written out and cut. The Mac catches soft and sentence-start fillers reliably and says plainly that some plain ums need a manual trim.
Free, private, undoable
Transcription and cutting run on your device, cost nothing on any tier and undo in one step.
What you need
- macOS 14 Sonoma or later (Apple Silicon or Intel), or iPhone on iOS 18 or later
- Spoken English: the filler list is um, uh, like, you know and sentence starters such as so
- On iPhone, the transcription model downloads once (about 145 MB) the first time you use it
- Free on every tier: no watermark, exports up to 2 minutes at 1080p
Questions
Why are some ums still there on my Mac?
The Mac transcribes with Apple's speech engine, which often leaves plain um and uh out of the transcript, so the filler pass cannot see them. Soft and sentence-start fillers are caught reliably. Cut leftover ums by hand, or do this job on iPhone, where the transcript is verbatim.
Does it work offline on iPhone?
Yes, after a one-time download of the transcription model, about 145 MB. If it has not downloaded yet, Zella falls back to Apple's on-device recognizer, and a few stray ums may need a manual trim.
Which languages does filler removal handle?
It looks for English fillers: um, uh, like, you know, and sentence starters such as so or I mean. Zella captions 27 spoken languages, but on the Mac a take in another language gets its silences cut and skips the filler pass.
Is it free?
Yes. Filler removal, silence removal and Auto-Polish are free on every tier on Mac and iPhone, with no watermark.
Related tools
Read the manual