The filler hunt stops taking your evening.
AVTrim is a macOS app. Upload a take and it comes back transcribed, every um, filler phrase and repeat already marked by AI-assisted detection. Drop the marks on the words that should stay. Nothing is cut until you say so. 7 day free trial, 3 audio-hours, email and a 6-digit code, no card. The build is unsigned, so the first launch needs Open Anyway.
have questions. Okay, hey, hey, welcome to the party YouTube. Welcome
to the party. Everybody is actually done with the
question YouTube. Okay, so I do have a lot to say about
that. I do think that there's a whole bunch to that that
makes a lot of sense.
There's two types of easy buttons.
There's easy button number one, which is I don't know how to
spawn a process from JavaScript or TypeScript or bun
or whatever. And I want to spawn a process and be able
to listen to the standard out and standard error and get
the exit code when it's done. Can you please give me an
example of that? That is a huge easy button.
Like we have to all agree that that is
one of the the greatest things ever. You only have some
You decide what counts as a filler in your own take.
The transcript comes back in four lists: um and uh, the words you added yourself, the softer fillers AI-assisted detection picks up ("i mean", "kind of", "you know"), and the words you repeated. Switch a whole list off, or leave it on. If "you know" is part of how you talk, it stays. Take the rest out of the audio or mute them, whichever suits the project.
You see how many times you actually said it before you switch it off.
Nudge either edge by a millisecond until the cut sounds right.
Your next recording starts where the last one left off.
Speakers are identified before analysis, and you name each one or link a profile you saved earlier. That profile remembers your mute-timing corrections for that voice, with AI-assisted detection on or off. Add your own filler words in the ADD NEW tab. The rest runs on keys: unmark, bookmark a line, jump to the next marked word. There are shortcut sheets for the transcript and for fine-tune, and view filters: clean only, repeats only, context only.
You can get through a whole pass without the mouse.
Render it here, or hand your editor the cut list.
Every step stays yours. The finding is the one you hand over.
| What you do | Doing it by hand | With AVTrim |
|---|---|---|
| Get the words down | You listen end to end | Transcribed as you uploadTranscribed on our servers, nothing to install |
| Find the fillers | You scrub the waveform | Marked before you open itUms, repeats, your own words, and soft fillers from AI-assisted detection |
| Decide what stays | You judge one cut at a time | You drop the mark, it staysAccept and move on with one key, or drop the mark on a word, a list or a speaker |
| Tidy the edges | You drag handles by eye | You nudge them by a millisecondExact in and out timecodes on screen while you listen |
| Get it out | You re-cut the timeline | You render it or send a cut listRender here, render on the cloud, or hand Premiere, Resolve or Final Cut a cut list |
| Who says what gets cut | You do | You do |
Work out what the filler hunt costs you in a year
Put your rate against the hours the hunt takes. Three numbers, all your own.
Three numbers you already know
Pay by the hour of recording you analyze.
Free trial
$0 for 7 days
3 audio-hours, no card
your own recording, not a sample
- Everything Studio has, for the whole 7 days
- Every filler type: um/uh, your words, soft fillers, repeats
- Mute or cut, switchable per project
- Hear every fix before you accept it
- Cloud Export and Pro Export included
Creator
$12/mo
$120 a year, two months free
$10 a month on the yearly plan
- 10 audio-hours a month
- Every filler type: um/uh, your words, soft fillers, repeats
- Mute or cut, switchable per project
- Fine-tune any boundary in 1 ms steps
- Keyboard shortcuts for the whole review
- Render locally, no video credits used
Studio
$24/mo
$240 a year, two months free
$20 a month on the yearly plan
- 30 audio-hours a month
- Everything in Creator, plus:
- Cloud Export on a server GPU, no second meter
- Pro Export: cut list to Final Cut, Premiere or Resolve
- The cut list route leaves your master untouched
Start with 3 free audio-hours over 7 days. An email address and a 6-digit code, no card. The cap is checked before a run starts, so a recording that does not fit runs nothing. Top-up hours are $3 an hour for import video and AI, and $3 an hour for export video and AI.
What people ask before the first upload
Put tonight's take through AVTrim
3 free audio-hours over 7 days, on your own recording. An email address and a 6-digit code, no card. macOS only for now, and the build is unsigned, so the first launch needs Open Anyway in Privacy and Security.