词频分析器

What word frequency analysis reveals

Counting how often each word appears in a piece of writing turns prose into data. The most frequent words expose what a text is really about, surface repetition you may not have noticed, and let you measure keyword density for search-engine optimization. This analyzer takes any block of text, tallies every word, ranks them from most to least common, and shows each one's raw count, its share of the whole, and a visual bar — all computed in your browser the moment you type or paste.

How the tool defines a "word"

Before counting, the analyzer extracts words with a pattern that keeps letters (including accented European letters like é, ñ, and ü) and apostrophes, while treating everything else — spaces, digits, punctuation — as a separator. So don't stays a single word, but well-being becomes two words, well and being, because the hyphen is a boundary. By default it also lowercases everything, so "The", "the", and "THE" count as one word rather than three. You can turn that case-insensitivity off if capitalization matters for your analysis.

The percentage is calculated against every word

This is the subtle point that makes the output trustworthy. The percentage shown next to each word is its count divided by the total number of words in the text, including any stop words — not divided by the filtered list. So when you hide stop words, the visible percentages still reflect each word's true share of the whole document. A keyword at "2.5%" genuinely appears in one of every forty words you wrote, which is exactly the figure SEO guidance refers to. The unique-word and total-word counts in the stats line are computed the same honest way.

A worked example

Paste this sentence: "The cat sat on the mat. The cat purred." That is nine words total. With case-insensitivity on and stop words hidden, "the" (a stop word) disappears from the table, but it still counts toward the denominator. "cat" appears twice, so its frequency is 2 ÷ 9 = 22.22%, and it tops the list with the longest bar. "sat", "mat", and "purred" each appear once at 11.11%. The bar lengths are scaled so the most frequent visible word fills the bar fully and the rest are drawn in proportion to it.

Stop words: focusing on meaning

Stop words are the grammatical glue of a language — the, a, is, of, and, to, and dozens more — that appear constantly but carry little topical meaning. Left in, they dominate any frequency list and bury the words you actually care about. The built-in filter removes a curated set of these so the ranking surfaces content words. Toggle the filter off when you specifically want to study function-word usage, for example in stylometry or readability work.

Practical use cases

• SEO keyword density. Paste a page's copy, hide stop words, and check that your target keyword sits in the commonly recommended 1–3% range — high enough to signal relevance, low enough to avoid looking like keyword stuffing.
• Editing for repetition. A content word that ranks far higher than you expected is a sign you have leaned on it too much; the table tells you to vary your phrasing.
• Summarizing unfamiliar text. The top ten content words give a quick sense of a long document's subject before you read it.
• Research and corpus work. Export the full ranked list as CSV to analyze in a spreadsheet or notebook.

Common mistakes

• Chasing a higher density number. Modern search engines reward natural writing, not a specific percentage. Treat density as a sanity check, not a target to game.
• Forgetting hyphens split words. Compound terms joined by hyphens are counted as separate words, so "state-of-the-art" will not appear as one entry.
• Comparing percentages across the filter toggle. The denominator (total words) is stable, but the visible list changes when you hide stop words; read the stats line to keep totals straight.
• Ignoring case for proper nouns. With case-insensitivity on, "Apple" the company and "apple" the fruit merge into one count.

Privacy

All analysis runs locally in your browser. Your text is never uploaded or stored on a server, and the CSV export is generated on your device — so you can safely analyze private drafts, client copy, or unpublished work.