Word frequency counter

What a word frequency counter counts

A word count tells you how long a draft is. A word frequency counter tells you what the draft keeps saying. Paste the text and every distinct word is listed from most common to least common, with its count and its share of the words that remain after your filters. The table updates while you type and downloads as CSV, so the result can be sorted, charted, or compared with another version in a spreadsheet.

Capitalization is folded deliberately: Draft, draft, and DRAFT belong to one row. Apostrophes inside a word stay attached, so don't is one word rather than don plus t. Letters outside English are words too; Västerås is not cut into ASCII fragments merely because it carries diacritics.

When to ignore common English words

Words such as the, and, of, and to usually rise to the top of English prose because grammar needs them, not because the subject does. Turn on the stop-word option when you want topic terms and repeated phrasing to stand out. Leave it off when the exact linguistic distribution matters, or when the text is not English. The list is intentionally short: the tool should not silently decide that an ordinary content word is unimportant.

Minimum length is a separate filter. Setting it to three or four characters removes initials, one-letter variables, and tiny fragments without applying any language judgment. Both filters change the denominator used for the percentage column, so each share describes the table you are actually looking at rather than the hidden words.

What a high count can and cannot tell you

Repetition is evidence, not a diagnosis. A product name may need to appear twenty times in a manual. A filler phrase appearing twenty times in an essay may need an edit. The useful workflow is to notice a surprising row, search for that word in the source, and read the sentences around it. Replacing every frequent word with a synonym usually makes writing less precise rather than more varied.

The same caution applies to SEO. Frequency is not a keyword-density target, and the table does not score or optimize a page. It shows literal occurrences so you can inspect your own writing. Search engines understand inflection and context; forcing a phrase to reach a percentage is an excellent way to make a page worse for the person who has to read it.

Clean the input only when the mess is not meaningful

Text copied from a PDF or spreadsheet can bring line breaks, invisible spacing, and repeated headers. Those characters do not become words here, but repeated header text does. If the source is visibly damaged, run it through the text cleaner first. If the repetition consists of whole duplicated rows rather than vocabulary, removing duplicate lines is the more precise operation.

Nothing is uploaded. The text, the full frequency table, and the downloaded CSV are produced inside this tab. That matters for interview transcripts, unpublished drafts, customer feedback, and every other document whose repeated words are interesting precisely because the document is not public yet.