Filler and empty-word analyser
This tool measures how much of your text is taken up by function words, filler intensifiers, officialese and cliches, and shows every match on its own line — the word, its category and a hint on what to put there instead. If you are starting out, you need this for one thing: to see what to edit. The general advice "write shorter" gives you nothing; a list of twenty specific words and the three densest sentences gives you a place to start. The text is processed in your browser and is never sent anywhere.
Text to check
Your text is never sent anywhere: the dictionaries are built into the page and the analysis runs locally.
Filler index
Match categories
All matches
| Word or phrase | Category | Times |
|---|---|---|
| Paste your text and the matches will appear here. | ||
The three densest sentences
The sentences with the highest concentration of matches will appear here in full, with every match highlighted.
How the filler index is calculated
The filler index is the share of words from four categories in the total word count: function words (prepositions, conjunctions, pronouns, particles), filler intensifiers, officialese and cliche openers. The formula is simple: those words divided by all the words in the text. Passive voice and repeats are counted separately and stay out of the percentage. They do not thin the text out, they make it impersonal: "it was decided" does not say who decided, and the same stem three times in forty words reads as running in place. You need a benchmark. Even in carefully edited English prose function words account for 35 to 45 per cent: prepositions and conjunctions have to go somewhere. So "zero per cent" is neither a goal nor an achievable result. Read the list of matches, not only the number.
Six categories and what to do about each
Filler intensifiers are "very", "really", "actually", "basically", "literally". They come out at no cost: instead of "a very large print run" write the size of the print run. Officialese covers "utilize", "due to the fact that", "in the event that", "for the purpose of", "at this point in time". Each of these folds into one verb or one preposition: "utilize" becomes "use", "due to the fact that" becomes "because". Cliche openers are "it is well known", "in this day and age", "needless to say", "we all know". They are a run-up before the thought. The reader came for the thought, so the run-up is cut whole. Passive and impersonal forms are "it is considered", "it was decided", "a decision was made". Give the action back to whoever acts. Repeats are the same stem three times inside a forty-word window; use a synonym or rebuild the sentence. Function words are always needed, but long chains of prepositions are worth cutting apart.
Why whole sentences are shown
A list of words tells you what is wrong. A sentence tells you where. So the tool picks the three sentences with the highest density of matches — flagged words divided by all the words in the sentence — and shows them in full, with every match highlighted. This is the fastest way to edit: you get a ready piece of text in which everything removable is already marked. Those three sentences usually get cut in half in about a minute.
What the tool does not do
It does not understand meaning. The dictionaries are built into the page and there is no morphological analysis: passive voice and repeats are detected by endings and stems, so both misses and false hits happen. The word "just" in "a just decision" is not filler, and the tool may still flag it. It does not judge the quality of your writing. A low percentage does not make an article interesting, and a high one does not make it bad: fiction dialogue and a legal document have different norms. And the main point: if the checks found nothing, that only means the built-in dictionaries stayed quiet. It is not a verdict that the text is ready to publish — that verdict comes from an editor, not from a program.