Remove Filler Words from Dictation Automatically — Um, Uh, No More

Everyone says "um" and "uh" when they talk. That's fine in conversation. It's annoying in transcribed text. Whisperer can strip these out automatically.

How to enable it#

It's off by default (sometimes you want exact transcription). To turn it on:

1. Open Settings
2. Find "Filler Word Removal" under Text Processing
3. Toggle it on

Done. Filler words get stripped from all future dictation.

What gets removed#

These words: um, uh, erm, er, ah, hmm.

The processing pipeline#

Filler removal is one step in a longer chain. Every transcription goes through these stages:

1. Transcription — Whisper/NVIDIA/Apple Speech produces raw text
2. Dictionary corrections — fixes your custom terms (details)
3. Filler word removal — strips um, uh, etc. (optional)
4. List formatting — converts spoken lists to bullets (optional)
5. AI post-processing — rewrite, translate, format (details) (optional)
6. Trailing space — adds a space so you're ready to keep typing (optional)
7. Injection — text goes into the focused field

Each step can be toggled independently. Mix and match.

List formatting#

Whisperer can turn spoken lists into formatted bullet points. No AI needed for this one.

It recognizes:

• Ordinals: "first", "second", "third"
• Cardinals: "one", "two", "three"
• Digits: "1", "2", "3"
• Phrases: "number one", "bullet point", "dash"

There's false-positive prevention built in. "I had two coffees" doesn't become a list. "About three hours" stays as prose. It checks 60+ preceding words to avoid mistakes.

If the rule-based engine misses something, there's an optional AI fallback that looks for implicit list structure.

When to use what#

Getting started#

1. Download Whisperer
2. Turn on Filler Word Removal in Settings
3. Optionally enable List Formatting
4. Try AI post-processing for more cleanup options