Filler word counter how many um, like and you know you say.
Paste a speech transcript or open an SRT or VTT subtitle file to see how often you say um, uh, like, you know, I mean, so, basically, actually and right. With the subtitle timing, the counter also shows fillers per minute and your pace in words per minute, and highlights every filler.
Free, no sign-upResult
- Fillers found
- 0
- Per 100 words
- 0
- Words
- 0
- Fillers per minute
- -
- Words per minute
- -
Add the duration (or open a timed subtitle file) to see the per-minute numbers.
Most used
The fillers found show up here, from most to least used.
Where they show up
Every filler counted is highlighted in yellow.
Want a transcript of your talk? We transcribe the whole video. Try it free
How to count the filler words in a speech
- 01
Paste the transcript or open the subtitles
Plain text, an SRT, VTT or SBV subtitle, or a transcript JSON all work. Timestamps, cue numbers and speaker labels are removed before counting.
- 02
Check the language and duration
The language is detected from the text. The duration comes from the subtitles, or you type it, to get fillers per minute and words per minute.
- 03
Read the numbers and the highlights
The list goes from your most used filler to the least used, and the text shows each one highlighted. Add your own words and copy the result.
What are filler words?
Filler words are the words and sounds we use to buy time while we think or to keep a conversation moving: um, uh, like, you know, I mean, so, basically, actually, right. Everyone uses them, in every language. In conversation they sound natural and even help listeners follow along.
They become a problem when one turns into a habit: the same "you know" closing every sentence of a lecture, "so" opening every point of a talk, "like" three times in one line of a video. The audience starts counting instead of listening. This counter is there to help you see your own pattern, not to drive the list to zero.
How does the counter know when "like" is a filler?
Many of these words also have a normal meaning. "I like this light" has no filler; "it was, like, huge" does. So each word has a context rule, and when in doubt the counter would rather miss a filler than flag a word used correctly.
Hesitation sounds (um, uh, erm, hmm) and basically, actually and literally always count. "So" counts only when it opens a sentence, and "right" only as a tag ("it's done, right?") or as an opener ("Right, let's go").
| Word | Counts when | Not in |
|---|---|---|
| like | set off by commas, as in "she was like, wow" | I like it, looks like rain |
| you know | used as a tag | do you know him, you know what |
| I mean | used as a reset | that's what I mean, I mean it |
| so | it opens a sentence | I was so tired |
| right | a tag question or an opener | turn right here |
How many filler words per minute is too many?
There is no magic number, and be wary of anyone who gives you one. A podcast conversation carries far more fillers than a prepared keynote, and an edited short video carries fewer still. The useful way to use the counter is to compare you with you: record, count, practice and count again on the next recording.
Two measures let you compare recordings of different lengths. Per 100 words needs no timing and works with any text. Fillers per minute and words per minute need the duration, which comes straight from an SRT or VTT file (from the start of the first cue to the end of the last) or which you type in.
Your pace in words per minute tells part of the story too. Rushed speech tends to carry more hesitation; speech with pauses, less. For reference, conversation sits around 150 words per minute.
Can it count fillers in Portuguese and Spanish?
Yes. The language is detected from the text, and you can switch it by hand. In Portuguese we count né, tipo, então, aí, assim, sabe, entendeu, cara, é... and the ahn and hum sounds. In Spanish, este..., o sea, pues, bueno, eh, mmm, como que, digamos, ¿no? and vale, each with its own context rule.
If you have a habit that is not on the list, like "kind of", "okay" or "honestly", type it in the other words field. Those count as whole words anywhere in the text and show up in the list and the highlights.
More than a filler counter. Your whole talk, in text.
Frequently asked questions
How do I know how many times I say um?
Transcribe your talk, then paste the text here or open the SRT file. The counter shows how many times um appears, along with the other fillers, and highlights each one. With the duration, you also see how many per minute.
Are filler words bad?
No. They are part of everyone's speech and keep conversation natural. What gets in the way is repetition: when the same filler shows up in almost every sentence, it draws more attention than your content. The goal is to notice the pattern and cut back, not to sound like a robot.
Why didn't it count every "like" I said?
Because like is also a normal word, as in "I like it" or "looks like rain". The counter skips it after words like I, would, looks and feels, and before words like this or a unless commas set it off. To count every single one, type like in the other words field.
How do I calculate speaking rate in words per minute?
Divide the number of words by the minutes of speech. Open the recording's SRT or VTT file and the counter reads the duration for you; with pasted text, type the duration as 12:30. The result shows in the words per minute box.
Does it work with a YouTube transcript?
Yes. Paste the transcript text: timestamps and line breaks are removed before counting. If you download the subtitles as SRT or VTT, the duration comes along and you get fillers per minute too.
Is the filler word counter free?
Yes. It is free, with no sign-up and no usage limit. A FilmScribe plan only comes in if you want us to transcribe the full recording.
Your whole talk in text, to review at your own pace. With the time of every sentence.
Upload the video of your class, sermon or podcast. We transcribe it word for word, with who spoke and when, and you download TXT, PDF or SRT to count here.
See the result on your own recording before adding a card.