How to Remove Um and Uh From Your Podcast on Mac

Updated: May 2026

Filler words — um, uh, like, you know — are a natural part of speech. But in a finished podcast, they fragment your delivery and distract listeners. Removing them manually means hours of scrubbing through a waveform in a DAW. AudioClean Pro automates the entire process on your Mac, using on-device AI to find and cut every filler in minutes.

Why filler words hurt podcast quality

Research on listener retention consistently shows that verbal clutter erodes confidence in the speaker and makes content harder to follow. A podcast full of "um" and "uh" signals hesitation even when the underlying ideas are strong. Removing fillers:

  • Makes your delivery sound more polished and authoritative
  • Reduces episode length — a 45-minute interview often drops to 40 after filler removal
  • Improves transcript readability for show notes and SEO
  • Saves listener patience, especially on long-form episodes

The challenge has always been time. Manual removal in Audacity or a DAW requires you to find each filler visually or by ear — a process that takes as long as the episode itself. AudioClean Pro replaces that with a single pass.

How AudioClean Pro removes filler words

AudioClean Pro uses WhisperKit — Apple's on-device transcription framework — to transcribe your recording directly on your Mac. No audio is sent to a cloud server. The transcription engine identifies every word, including fillers, with timestamps accurate to the millisecond.

Once the transcript is ready, AudioClean Pro highlights all detected filler words in a review editor. You see each filler in context — the sentence around it — so you can decide whether to remove it or leave it. Removing a filler that anchors a sentence ("like, this is the thing") is easy to catch at the review step.

When you approve and export, AudioClean Pro splices out each approved filler and smooths the edit so there are no click artifacts or awkward gaps. The result sounds like a natural, confident delivery — not like something has been cut.

Step-by-step: removing um and uh on Mac

  1. Open AudioClean Pro and drag your podcast recording into the app. Supported formats: MP3, WAV, FLAC, M4A, MP4, MOV.
  2. Enable Filler Word Removal in the processing panel. You can combine this with noise removal and silence trimming in the same pass.
  3. Wait for transcription — typically 1–3 minutes per hour of audio on Apple Silicon. Everything runs locally; no internet connection needed.
  4. Review the transcript. AudioClean Pro highlights each detected filler. Uncheck any you want to keep; the rest are queued for removal.
  5. Export. Choose your output format (MP3, WAV, FLAC, M4A). AudioClean Pro writes the cleaned file — original is never modified.

The entire workflow, including review, typically takes 5–10 minutes for a 45-minute episode.

What filler words are detected

AudioClean Pro detects a broad set of English filler words and hesitations:

  • Pure hesitations: um, uh, er, ah
  • Softeners and qualifiers: like, you know, sort of, kind of, basically, literally
  • Discourse markers: right?, okay?, so (when used as a filler opener)

The transcript editor lets you see each detected filler in context, so over-detection (catching "like" when used meaningfully) is easy to correct in the review step.

Combining filler removal with noise removal

For podcast recordings, filler word removal and noise removal often need to happen together. A home studio recording might have background hum, room reverb, and frequent filler words all at once. AudioClean Pro handles all of these in one processing pass:

  • Noise removal — eliminates background hum, fan noise, air conditioning, street noise
  • Filler word removal — cuts um, uh, and hesitations automatically
  • Stutter removal — smooths word-level repetitions ("I-I-I think")
  • Long silence trimming — reduces awkward pauses longer than a threshold you set

Running all four in a single export avoids the quality loss from re-encoding multiple times.

AudioClean Pro vs manual filler removal in Audacity

Method Time per 45-min episode Accuracy Privacy
Manual (Audacity / DAW) 2–4 hours High (human judgment) On-device
Cloud tools (Cleanvoice, Descript) 10–30 min + upload wait Good Files uploaded to cloud
AudioClean Pro 5–10 min High (AI + human review) 100% on-device

Pricing

AudioClean Pro is available at a launch price of $59 (regular $89) — a one-time purchase from the Mac App Store. No subscription, no monthly fees, no per-minute credits. A 14-day free trial with full feature access lets you run your complete workflow before buying.

Download AudioClean Pro for Mac →

AudioClean Pro home · Pricing · For podcasters · Alternatives · Why fillers hurt listeners