Word Frequency Counter: A Step-by-Step Guide
Reviewed by the FreeOnline.fyi team · Updated 2026-09-16
Paste text and counts appear
The whole interface is one wide box at the top of the page. Open the Word Frequency Counter, paste anything — a draft article, an essay, a lecture transcript, an .srt subtitle file, a pile of customer reviews — and the analysis runs as you type. There is no submit button, no account, and no upload step.
The textarea accepts up to roughly 2,000,000 characters, which is more than most novels. A few hundred kilobytes of plain text is near-instant on a normal laptop; extremely long inputs may take a moment, because the browser has to tokenise every word before drawing the table.
The results panel sits directly beneath the box and reads something like 'Top 20 of 1,842 words · 604 unique words'. That single line already tells you two useful things: how long the sample is, and how repetitive it is.
Two switches change everything
Beneath the textarea sit just two controls, and between them they decide almost everything about the output.
The first is the matching selector. Case-insensitive is the default, so 'The', 'THE' and 'the' collapse into one row. That is usually right for prose, because sentence-initial capitalisation would otherwise split a common word into two or three separate rows.
Switch to case-sensitive when capitalisation carries meaning: product names such as 'Apple' versus 'apple', code identifiers, or poetry where a capital letter is deliberate. Expect the ranking to fragment — the counts get spread across more rows, and the top-20 list looks sparser than the text really is.
The second control is the 'ignore common English stopwords' checkbox, on by default. It strips high-frequency function words like the, and, of and to, which is what makes the ranking readable. That idea is standard in text analysis; the Wikipedia entry on stop words covers its history. Uncheck it when those words are the point — comparing two writers' sentence rhythm, or checking how often a translation leans on 'that'.
Reading the top-20 table
Each row shows a rank, the word itself, a raw count, and a percentage. The percentage is the word's share of all counted words — that is, after stopwords have been removed if that checkbox is on. The twenty rows do not add up to 100% unless your text contains twenty distinct words or fewer.
A slim horizontal bar behind each row is scaled to the most frequent word, so the drop-off is visible at a glance. In ordinary English the second word typically lands near half the count of the first, and the tenth is a small fraction of it — a long-documented pattern known as Zipf's law.
If your text has fewer than 20 distinct words, empty rows are simply omitted rather than displayed as zeroes, so the table is never padded with noise.
Export the list to CSV
The 'Download CSV' button produces a file named word-frequency.csv with four columns: rank, word, count, percentage. You can export the full ranked list or just the visible top 20, depending on what you need downstream.
Two details make the file safe to open in a spreadsheet. Fields containing commas, quotes or line breaks are quoted and escaped, so an entry like 'state-of-the-art, reviewed' cannot shift your columns. A UTF-8 byte-order mark is also prepended, which is why Excel renders accented and non-Latin words correctly instead of turning them into mojibake.
The file is assembled in your browser from a Blob and an object URL, so nothing is posted anywhere. Once it is open, the usual next moves are a pivot table, a scatter chart of count against rank, or a lookup against your brand's preferred-terms glossary to see which words you actually overuse.
What the counts can't tell you
A frequency list is a lens, not a verdict. The counter compares exact strings, so run, running and runs occupy three separate rows — there is no stemming. Hyphenated compounds, contractions such as don't, and numeric tokens each count the way they are written. Decide in advance whether that matters for your question.
The bigger trap is reading the top of the list as a summary. The most repeated word in a legal brief is often a party's name; in a recipe post it is often an ingredient mentioned once in the title and then repeated throughout. Neither tells you what the document is about in any interesting sense.
Use the counter alongside reading, not instead of it. If a number will drive a decision — keyword research, a content audit, a claim in a paper — check it against the source text before you act on it.
Privacy, limits and other tools
Everything runs client-side: your text is tokenised in the browser and the CSV is built locally. Nothing is sent to a server, which is the reason pasting confidential material here is reasonable in the first place — and the reason the page can stay free and account-free.
Very large pastes can make any tab sluggish, so if you are analysing a whole book, consider splitting it into chapters and comparing the exports afterwards. And sanity-check surprising results: an unexpectedly low count usually means the case setting or the stopword filter is doing more than you realised.
The same site hosts a growing set of free online tools built on the same no-signup, browser-first pattern, from text utilities to planning helpers.