Filler-word removal
EasyCast finds filler words in a recording’s word-timed transcript and cuts them out of the audio, so um, uh, and phrases like “you know” stop slowing the episode down. While the edit is still the recording’s most recent change, Undo on the recording page restores the audio and transcript from just before it.
Reviewed by EasyCast Studio - Updated October 4, 2026
Works on
Audio recordings
Included on
Pro & Business
Reversible
Undo the latest edit
Short answer
Yes. Once a recording has a word-timed transcript, EasyCast queues an audio edit that trims the filler words it found, renders an MP3, and saves the shortened transcript alongside it. If no filler words match, the original recording is left unchanged. An applied edit can be undone from the recording page while it is still the recording’s latest change; Undo restores the snapshot from just before that edit, and is rejected once a later edit or another change to the audio or transcript supersedes it.
What the feature does
EasyCast is built for spoken audio, so filler removal reads the transcript’s word timing and edits the recording rather than stopping at the text.
The recording page’s Remove fillers action posts to /api/remove-fillers, which builds an audio cut list from the filler words found in the transcript, renders the shortened audio, and saves the transcript with those words removed and the remaining timestamps shifted.
Detection matches a defined list: um, uh, er, ah, basically, actually, literally, honestly; the phrases you know what I mean, you know, I mean, sort of, kind of, and you see; and immediately repeated words (stutters).
Filler detection reads the transcript’s word timing, so the recording must be transcribed first. Without a timed transcript the action stops and asks you to transcribe.
Undo posts to /api/undo-audio-processing and restores the audio and transcript snapshot from just before that edit, so a filler cut is not a one-way change while it remains the recording’s latest edit.
The audio edit is offered for recordings whose file type is audio; a video recording returns an audio-only message instead of queuing a job.
If the transcript contains none of the listed fillers, EasyCast reports no changes and leaves the original recording in place.
Workflow map
Filler removal is one step inside the recording page, next to transcription and silence removal.
| Step | Job | EasyCast fit |
|---|---|---|
| 1 | Start with an audio recording | Filler removal is available for recordings whose file type is audio. |
| 2 | Transcribe with word timing | EasyCast reads the transcript’s word timing, so give the recording a word-timed transcript first. |
| 3 | Run Remove fillers | The action finds the listed filler words in the transcript and queues an audio edit that cuts them. |
| 4 | Review the edited audio | When the job finishes, the recording carries the edited audio, rendered as an MP3. |
| 5 | Undo if the take changed too much | Undo restores the audio and transcript snapshot from just before the edit, and is rejected once a later edit or another change to the audio or transcript supersedes it. |
Decision guide
Use Remove fillers
The list targets the common English filler sounds, the listed phrases, and immediate repeated words.
Plan on a manual edit
Those words are not on the fixed EasyCast filler list, so a single use is left in; an immediate repetition such as “like like” is still cut as a stutter even though the word itself is not listed.
Transcribe first
Filler removal needs word timing, so it is unavailable until the recording has a word-timed transcript.
Use Undo after the edit
While the edit is still the recording’s latest change, Undo restores the audio and transcript from just before it, so you can reverse a cut that changed the meaning.
Pro or Business
Filler-word removal is included on Pro ($29/month) and Business ($99/month). It is not included on Free or Starter. Starter includes transcription and basic enhancement but not filler removal; choose Pro when you need it.
Why EasyCast
EasyCast keeps transcription, filler removal, and silence removal on the same recording page, so a cleanup pass does not mean exporting into another tool.
Finds fillers from a word-timed transcript, not guesswork
Cuts the audio and keeps an Undo on the recording page
Runs beside transcription and silence removal in one workspace
Filler removal is included on Pro ($29/month) and Business ($99/month)
Compare all-in-one podcast software against specialist recorders, editors, cleanup tools, and hosts before you commit your workflow.
See the broader AI editing workflowClean up podcast audioCompare podcast editing softwareGet a word-timed transcriptFAQ
Both the audio and the transcript. Remove fillers queues an audio edit that cuts the listed filler words out of the recording and renders an MP3; the same edit removes those words from the word-timed transcript and shifts the remaining timestamps to match the shorter audio, and the updated transcript is saved with the recording.
A fixed English list: um, uh, er, ah, basically, actually, literally, honestly; the phrases you know what I mean, you know, I mean, sort of, kind of, and you see; and immediately repeated words. A word that is not on that list — like “like” or “right” — is left in on its own; an immediate repetition such as “like like” is still cut as a stutter.
Yes. Filler removal reads the recording’s word timing, so it needs a word-timed transcript; without one the action stops and asks you to transcribe the recording first. EasyCast transcription is English only.
Yes, while that edit is still the recording’s latest change. The recording page keeps an Undo for the last audio edit; it restores the audio and transcript snapshot from just before the edit, and is rejected once a later edit or another change to the audio or transcript has superseded it.
Filler-word removal is included on Pro ($29/month) and Business ($99/month). It is not included on Free or Starter. Starter includes transcription and basic enhancement but not filler removal; choose Pro when you need it.
No. The audio edit is offered for recordings whose file type is audio; a video recording returns an audio-only message.
Try the workflow
Start with EasyCast if you want recording, AI production, publishing, and conversion learning closer together.
