Word Frequency: Online Word Count Analyzer and Word Cloud

The tool counts how many times each word occurs in your text and shows the result in two ways: as a word cloud, where the size of a word is proportional to its frequency, and as a table with exact occurrence counts and shares. Paste text into the input field — the analysis runs as you type. The cloud can be downloaded as PNG, and the table copied or exported as CSV.

A bigram is a pair of adjacent words, such as "search engine optimization" or "product page". Single words show what a text is about; bigrams show the wording it uses to say it. For SEO that is closer to real search behaviour, because people search in phrases rather than in isolated words. The switch above the settings changes the mode for the cloud and the table at once.

Stop words are the words you do not want to see in the result. Most often they are function words: prepositions, conjunctions, pronouns, particles. They are the most frequent words in any text but say nothing about its content — without them the top of the table fills up with "the", "of", "and", "to". The tool does not impose a ready-made dictionary: what counts as redundant depends on the text and the task, so you set the list yourself. Type everything unnecessary into the "Stop-word list" field — function words, a brand name, SKUs, recurring phrasings. Any separator works: a comma, a space, a line break. The list applies immediately and stays fully visible, so it is always clear what exactly was left out of the count.

It is the ratio of unique word forms to the total number of words, expressed as a percentage — in linguistics this is the type-token ratio. A hundred percent means no word was repeated. The lower the value, the more the text circles around the same words. The measure depends on length: a short text always scores higher, so only texts of comparable size are worth comparing.

From an average silent reading speed of 200 words per minute, rounded up to a whole minute. It is a rough guide to length rather than a precise measurement: actual speed depends on the complexity of the text, the language and the reader. Technical prose full of terms and formulas reads noticeably slower than fiction.

It cuts out short function words and fragments that carry no meaning, and it does so without a stop-word list — handy when listing them one by one would take a while. The default is 3 characters. Raise it to 5–6 to keep only meaningful vocabulary, or lower it to 1 to see every token, numbers included. The limit does not apply to Chinese characters or Hangul: a Chinese or Korean word is one or two signs, and a shared threshold would wipe them out entirely.

The page shows the first ten entries — enough to see the picture without stretching the page. The copy and CSV buttons give you the entire list, however long it is. The line below the table always states how many rows are shown and how many there are in total.

Nowhere: the whole analysis runs in your browser, on your device. The text is never sent to a server, stored or written to any log, so drafts, internal documents and client materials are safe to run through it. The image and the CSV file are generated locally as well.