Filler word counter for any transcript.

What this filler word counter looks for

Filler words are not one thing, so the tool sorts what it finds into nine groups rather than lumping them into a single number. Hesitations are the classic ones: um, uh, er, ah. Discourse markers are the words that pad a sentence without adding to it, like, you know, basically, actually, I mean. Filler phrases are the longer versions, such as at the end of the day or to be honest with you.

The other six groups are about writing rather than speaking. Clichés are phrases that have stopped carrying meaning. Wordy phrases are the ones with a shorter equivalent, like in order to. Passive voice, run-on sentences, weak openers such as there is and it is, and weak intensifiers such as very and really all get their own colour.

Each match is highlighted in place and counted in a breakdown, so you can filter to one category and see only that. That matters because the fix differs by type. You delete a hesitation. You rewrite a passive sentence.

How the speech grade is worked out

The grade is the share of your words that were flagged, not a raw count. A hundred filler words in a ten thousand word transcript is a different situation from a hundred in a thousand, and a raw count would rank them the same.

Under about one and a half per cent scores an A and reads as exceptionally clean. The grade slides down from there. Short samples do not get a grade at all, because a rate calculated over twenty words tells you nothing reliable.

Treat the grade as a direction rather than a verdict. Spoken English is naturally full of hesitations, and a transcript that scores worse than an essay is not a worse piece of thinking. It is just a transcript.

Private filler word counting, with nothing uploaded and nothing to install

Your text stays in the page. The matching happens in JavaScript in the tab in front of you, so nothing is posted anywhere and there is no account to make.

This tool is also different from the other Yaps tools in one useful way: there is no model to download. The speech and image tools fetch a model file on first use, which takes a moment. This one is a dictionary and a set of rules, so it works the instant the page loads, and it keeps working with the network off from that moment on.

That makes it safe for text you would not paste into a random website. Interview transcripts, client calls, medical or legal notes, unpublished drafts. If it would be awkward for that text to sit in someone else's logs, this is the kind of tool you want.

How to count the filler words in a transcript

Paste your text or drop a PDF, EPUB, TXT, Markdown, HTML, RTF or DOCX file into the box. Anything works: a transcript from a recording, a speech script, a blog draft, an email you are about to send. Files are parsed on your device and there is no length limit.

Read the summary. You get a total, a rate, a grade and a read time, plus a breakdown by category. Click a category to filter the highlights down to just that type.

Work through the highlights. Hesitations and discourse markers can usually just be deleted. Wordy phrases have a shorter form. Passive sentences need a subject. Weak openers usually mean the real subject of the sentence is buried later in it.

Switch to the cleaned view to see the text with the obvious fillers stripped out, then copy it. Treat that as a starting point rather than a final draft, because removing a word is not the same as rewriting a sentence.

Why filler words matter when you speak

Everyone uses them. They are a normal feature of unscripted speech, and a talk with none at all can sound recited. The problem is density and habit. A speaker who says like six times in a sentence is asking the audience to do the work of ignoring it.

They also cost time. In a recorded talk or a podcast, hesitations and restarts are dead air you pay for twice: once in the recording and again in the edit.

The useful part is not the count, it is the pattern. Most people lean on two or three crutch words rather than the whole list. Once you know which ones are yours, you can hear yourself doing it, and that is the point at which it starts to change.

Who uses a filler word counter

Speakers and presenters run their rehearsal transcripts through it to find the crutch words before an audience does. Podcasters and video editors use it to spot the passages that will need the most cutting.

Writers use it on drafts rather than transcripts, where the wordy phrases, weak openers and passive sentences matter more than the ums. Students and researchers use it to tighten an essay against a word limit.

It is also useful for anyone who dictates. Speech to text is faithful, which means it captures the false starts along with the sentences. Running the result through here shows you what your dictation habits actually look like.

What this tool does not do, and what the app does instead

This tool counts and highlights. It does not rewrite. The cleaned view removes obvious fillers, but it will not restructure a sentence, fix your punctuation or decide what you meant. That is deliberate, because a dictionary can find a passive sentence but it cannot rewrite one well.

It is also matching against fixed dictionaries and rules, so it is precise about what it knows and blind to what it does not. It will not catch a filler habit unique to you, and it can flag a phrase that is doing real work in your sentence. You are the editor.

The Yaps desktop app closes that gap at the other end. It runs a cleanup model on your machine while you dictate, so the filler words never reach the page: the ums and false starts are gone, punctuation and capitals are put in, and what lands at your cursor is already a clean sentence. This page tells you about the problem after the fact. The app removes it as you speak.