Filler Word Counter: Count Um, Uh, Like and More
A filler word counter scans a transcript for um, uh, er, like, you know, so, basically and similar words, then shows each one in context and a rate per 100 words. Paste what you said, read the highlighted lines and count the repeated words. It reads text only; it cannot hear your audio.
Paste what you said, including the hesitations.
This counter reads text only. It cannot hear your voice, check the transcript against a recording or measure pauses.
How the counter works
The counter matches whole words and whole phrases, not fragments: “um” is counted in “um,” but not inside “umbrella”. Multi-word fillers such as “you know”, “I mean”, “kind of” and “sort of” count once. Matches are case-insensitive.
Firm fillers are the terms above that are almost always fillers in speech. Like is marked possible because it also works as a preposition and a comparison, so it is counted but excluded from the headline rate. The rate is firm fillers ÷ words × 100.
Repeated words are found separately: two identical words with only whitespace between them, such as “the the”. The check ignores capitalisation. Everything runs in this browser tab; no transcript is uploaded.
Worked example
Press Try a sample to load this 21-word transcript: “So, um, I actually think that, you know, like, the the plan is basically fine, right? I mean, literally, sort of.”
It contains nine firm fillers (so, um, actually, you know, basically, right, I mean, literally, sort of), one possible filler (like) and one repeated pair (the the). Nine ÷ 21 × 100 = 42.9 firm fillers per 100 words.
A high number is not a character judgement. It is a list of places to slow down, breathe and choose the next word before speaking.
From text to voice
Counting fillers in a transcript is useful for editing. Measuring them in speech is different: a transcript hides pauses, and a transcription service may drop or add “um”. The speaking rate test covers pace, and the speech time calculator covers length.
To practise the real thing, read a true story at your level, put it away and retell it aloud. The methodology page explains how the product separates fillers from repetitions and pauses.
Questions and answers
What counts as a filler word here?
The counter looks for um, uh, er, you know, so, basically, actually, I mean, kind of, sort of, right and literally. It also counts like, but marks it as possible because like has ordinary uses too.
Is “like” always a filler?
No. “It looked like rain” is a comparison, not a filler. The counter marks every like as possible and leaves it out of the firm-filler rate, so you can review each one in context.
How is the rate per 100 words calculated?
Firm fillers ÷ total words × 100. A transcript with 9 firm fillers in 21 words has a rate of 42.9 per 100 words. Possible uses of like are reported separately and are not included.
Does this tool listen to my recording?
No. It reads text only. There is no microphone and no audio analysis. To have your spoken fillers measured, retell a true story aloud in SoulfulStories and read the transcript and analysis.
Why does a normal “so” count as a filler?
The count is deliberately literal: every standalone so matches, including grammatical ones. Treat the number as a list of moments to review, not a verdict about your English.
What are repeated words?
Two identical words next to each other, such as “the the” or “I I”. The counter lists each pair and how often it happens; it does not remove them from the transcript.
Keep going
A transcript cannot tell you how you sounded.
Typing “um” is not the same as hearing it. SoulfulStories records your retelling, transcribes it and measures the fillers, pauses and pace in the audio you actually produced.
Written by the SoulfulStories editorial team. Updated 26 October 2026. How we work.