Word Frequency Counter — Ranked List

Paste any block of text and every distinct word comes back ranked by how often it appears, alongside its share of the total token count. That share is density, calculated against all tokens in the document, not just the words that survive the stopword filter. The percentages stay meaningful when you toggle filtering on and off.

The Load Sample button fills the input with a 130-word passage about a quick brown fox and a dog, short enough to verify by hand and long enough to produce a useful ranking. After two passes through the sample, fox and dog each appear 6 times and sit at rank 1 and 2 with a density of 4.58% each. Switching the Include stopwords toggle adds the at 17 occurrences (12.98%) and reshuffles every rank below it.

Output can be copied as plain text or as CSV, or downloaded as word-frequency.csv containing all rows, not just the visible page. Nothing is uploaded; the entire analysis runs inside the browser tab.

Getting Your First Ranking

1. Paste text into the input field, or press Load Sample to start with the bundled passage.
2. The results table appears immediately, no button press needed. Words are ranked by frequency descending by default.
3. Use Min length to exclude very short tokens. Setting it to 4 drops words like fox and dog, leaving only longer terms.
4. Use Show top to cap the result set at 10, 25, 50, 100, 250, or 500 rows. The CSV download always exports the full uncapped list.
5. Click any column header — #, Word, Count, or Density — to re-sort. A second click reverses direction.
6. Switch to Bars view for a horizontal frequency chart; each bar scales to the highest-count word in the current filtered set.
7. Type into the Filter box inside the results header to narrow the ranked list without re-analysing the source text.

The Density Formula

Density is occurrences ÷ total tokens × 100, where total tokens counts every word in the raw input before any filtering. This matches the convention used by SEO density checkers. A keyword appearing 3 times in a 300-word article is 1.00%, regardless of how many stopwords those 300 words contain. Computing density against only the filtered pool would inflate every figure whenever common words are excluded, and the number would shift depending on which filter setting you chose.

Tricky Inputs and Edge Cases

• Hyphenated compounds like well-known split into two tokens: well and known. With Include stopwords off and Min length set to 4, only known survives.
• Contractions like don't become don and t after the apostrophe split. With the default minimum length of 2, t is dropped; don survives and will appear next to do as a separate entry.
• Numbers are treated as tokens. 2026 in a date-heavy document ranks alongside words; set Min length to 5 to push single years off the list while keeping most vocabulary.
• An empty input produces no table. The results panel only renders once text is present.
• At 500 rows with 10,000-word input the full analysis runs in under 16 ms on a mid-range device, because the frequency pass is a single O(n) loop over the token array.