Word frequency counter
Which words does this text lean on? Ranked counts with percentages, common filler words optional.
How to analyze a text
- Paste your text into the box. The ranked list appears immediately and re-ranks as you type or edit.
- Leave "ignore common words" checked to filter out roughly 50 English function words like the, and, and of. Uncheck it if you want the raw picture.
- Adjust minimum word length (1 to 4 characters, default 3) to hide short leftovers.
- Read the results: each word gets a count, a percentage of the filtered total, and a bar scaled against the most frequent word.
A worked example: catching a pet word
Paste a 2,000 word essay draft. The stats line might read "2,000 words total, 412 distinct after filters," and the ranked list tells you what the essay leans on. Suppose really sits at rank six with 21 appearances and 1.7 percent. No single sentence felt repetitive while writing, but one adverb carrying nearly two percent of your meaningful words is a tic, and now it has a number attached. Search your draft for it, cut two thirds, and re-paste here to confirm. This catch-and-cut loop is the single most useful thing the tool does, and it is why the filter defaults to on: with "the" and "and" dominating the top of an unfiltered list, your pet words hide below the fold.
How words are counted
The text is lowercased, so Apple and apple count together, then split into words: runs of letters including common accented characters, with apostrophes kept so that don't stays one word rather than splitting into don and t. Leading and trailing apostrophes are stripped, which keeps quoted words clean. Numbers and punctuation are ignored entirely. Words are then filtered by your minimum length and, optionally, the stop word list, and ranked by count with alphabetical order breaking ties. Percentages are computed against the filtered total, not the raw word count, so they answer "what share of the meaningful words is this?" The display shows the top 60 with a note about how many more exist, and the whole computation runs locally in your browser on every keystroke.
Honest limitations
There is no stemming or lemmatization: run, runs, and running count as three separate words, so a concept can be more frequent than any single row suggests. The stop word list is English only and fixed; you cannot add your own words to it, though setting minimum length to 4 approximates a stricter filter. Accented-letter support covers Western European languages, but scripts outside the Latin alphabet will not tokenize correctly. Percentages are rounded to one decimal, so tiny counts in a long document can display as 0%. And because everything recomputes on each keystroke, pasting a whole novel will make typing feel sluggish; for very long texts, paste chapters rather than the book.
Frequently asked questions
What counts as a word?
Letter sequences, case-insensitive, with apostrophes kept ("don't" is one word). Numbers and punctuation are ignored. "Apple" and "apple" count together.
What does the stop-word filter remove?
The ~50 most common English function words: the, and, of, to, a, in, and so on. Turning it on surfaces the words that actually characterize the text, which is the usual reason for counting.
Why would I count word frequency?
Writers use it to catch overused pet words, SEO editors check keyword balance, students verify essay requirements, and it is the fastest way to see what a long document is actually about.