Word Frequency Counter

About Word Frequency Counter

How Frequency Is Calculated

The tool splits text on whitespace, normalises each word (lowercase if case-insensitive mode is on), then counts occurrences. The percentage for each word is its count divided by total words, multiplied by 100.

Word	Count	Total Words	Frequency
design	5	200	2.50%
colour	3	200	1.50%
layout	2	200	1.00%

Filtering Options

Option	What It Does	When to Use
Case insensitive	Treats "Apple" and "apple" as the same word	Almost always - unless you specifically need to distinguish capitalised forms
Minimum word length	Excludes words shorter than N characters	Set to 3 to skip "a", "I", "an", "or" without using a full stop-word list
Exclude stop words	Removes common English function words (the, is, at, and, etc.)	When you want to see only substantive content words

Common English Stop Words

Stop words are function words that appear frequently in every text but carry little meaning on their own. Excluding them reveals the words that actually distinguish one text from another:

Category	Examples
Articles	a, an, the
Prepositions	in, on, at, to, for, with, from, by, of
Conjunctions	and, but, or, so, yet, nor
Pronouns	I, you, he, she, it, we, they, this, that
Auxiliary verbs	is, are, was, were, be, been, being, have, has, had
Common adverbs	not, very, also, just, then, now

SEO Keyword Density

Keyword density is the percentage of times a target keyword appears relative to total word count. While Google has moved beyond simple keyword density as a ranking factor, it remains a useful signal for content quality:

Density	Interpretation	Action
Under 0.5%	Keyword barely mentioned	Consider adding more natural mentions if it is your target term
0.5% - 2%	Natural range for most content	No changes needed - this is a healthy range
2% - 3%	Somewhat high - borderline	Check if the repetition sounds natural when read aloud
Over 3%	Keyword stuffing risk	Replace some instances with synonyms or related terms

A 1,000-word article targeting "running shoes" with a 1.5% density would mention the exact phrase about 15 times. That feels natural for a product review but would be excessive for a general fitness article.

Practical Applications

Use Case	What to Look For	Why It Matters
SEO content review	Target keyword frequency and density	Ensure target terms appear enough without over-optimising
Editing for repetition	Non-stop words with unusually high counts	Repeated words make prose feel monotonous
Academic writing	Key terms and their distribution	Check that core concepts are discussed throughout, not just in one section
Interview analysis	Frequently mentioned themes and topics	Identify what interviewees talk about most without manual counting
Competitor analysis	Compare word frequencies across competing articles	Find terms competitors emphasise that you may have missed
Translation QA	Compare source and target word frequencies	Large frequency differences may indicate missed or added content

For a quick overall count of words, characters, and reading time, the word counter provides those totals. For checking how readable your text is, the readability score tool grades your text by complexity. All processing runs in your browser - your text is never uploaded.

What Does a Typical Frequency Distribution Look Like?

Natural English text follows Zipf's law: the most common word appears about twice as often as the second most common, three times as often as the third, and so on. In the one-million-word Brown Corpus, "the" alone accounts for roughly 7% of every word in the text (69,971 occurrences), with "of" a distant second at about 3.5%. The top 100 words cover around 50% of everything written or spoken in English, and the top 1,000 cover about 80%.

This is why excluding stop words matters so much. Without a filter, almost every piece of English text produces the same boring frequency chart: the, of, and, to, a, in. Excluding stop words strips out that predictable background and surfaces the words that actually distinguish one text from another - the keywords a search engine or an editor actually cares about.

Rank	Word	Brown Corpus Share	Cumulative
1	the	6.9%	6.9%
2	of	3.5%	10.4%
3	and	2.8%	13.2%
4	to	2.6%	15.8%
5	a	2.3%	18.1%
10	he	~1.0%	~24%
100	(varies)	~0.1%	~50%
1,000	(varies)	~0.01%	~80%

If your text is in English and the top word is anything other than "the", "a", or "and", you've already got a strong signal that you're over-using a content term or that the sample is too short to behave normally.

How Many Words Do You Need for Reliable Results?

Short samples produce unreliable frequency data. A 50-word paragraph tells you almost nothing - the most frequent word might appear three times by pure chance. For content analysis, aim for at least 300 words before drawing conclusions, and 1,000+ for anything you plan to optimise for SEO.

Sample Size	Reliability	Good For
Under 100 words	Very low	Spot-checks only
100 - 300 words	Low	Rough repetition checks
300 - 1,000 words	Moderate	Blog post editing, email polish
1,000 - 3,000 words	Good	SEO keyword analysis, article QA
Over 3,000 words	High	Research, corpus linguistics, competitor analysis

How Does the Counter Decide What Counts as a Word?

Behind the scenes, the tool runs a simple token rule: a word is any sequence of letters, digits, or accented characters (A-Z, a-z, 0-9, and the Latin-1 Extended range for accents like é, ñ, ü), optionally with apostrophes inside for contractions. Leading and trailing apostrophes are trimmed, so 'hello' and hello count as the same word.

Worked example. Take the sentence: "The cat's toy isn't the same as the other cat's toy." With case-insensitive mode on and no filters, that's 11 tokens total. "the" appears 3 times (27.3%), "cat's" appears 2 times (18.2%), "toy" appears 2 times (18.2%), and "same", "as", "other", "isn't" each appear once (9.1%). Turn on "exclude stop words" and "the", "as", and "isn't" drop out - now "cat's" and "toy" tie for first at 28.6% each.

Because the tokeniser doesn't do stemming, cat and cats are separate entries. That's intentional - stemming introduces ambiguity (is "running" a verb or a noun?) and the correct form depends on your use case. For rigorous linguistic analysis, run the output through a dedicated stemmer or lemmatiser. For SEO and editing, the raw forms are what matters anyway - Google ranks "running shoes" and "running shoe" separately.

Common Mistakes When Using Frequency Data

Keyword density obsession is the biggest trap. Google stopped weighting raw keyword density as a ranking factor more than a decade ago, and current guidance from Google's Search Central emphasises topical coverage, entity relationships, and helpful content over term repetition. A 2% density target is a rough editorial check, not a ranking lever - hitting 2.0% exactly does nothing if the content is thin. Write naturally for the reader first, then use frequency analysis to spot accidental over-use.

Second mistake: forgetting that word frequency tells you what's there, not what's missing. If your article about "running shoes" never mentions "cushioning", "drop", or "pronation", the frequency chart can't warn you - it only lists what you did write. Pair frequency counts with a topic outline or competitor content analysis to catch gaps.

Third: treating case-sensitive counts as meaningful when they aren't. "Apple" (the company) and "apple" (the fruit) are genuinely different entities, but in 99% of general text you just want "apple" counted once. Leave case-insensitive mode on unless you have a specific reason to distinguish proper nouns.

For stripping duplicate lines before running an analysis, the remove duplicates tool is useful. For a different angle on text statistics, the character counter breaks text down by character type.

Sources

• Wikipedia - Zipf's Law (Brown Corpus frequencies)
• Piantadosi (2014) - Zipf's Word Frequency Law in Natural Language, NIH/PMC
• PLOS ONE - Zipf's Law for Word Frequencies: Word Forms vs Lemmas
• Google Search Central - Creating Helpful, Reliable, People-First Content
• Sketch Engine - Word List and Frequency Documentation
• NLTK Book - Language Processing and Frequency Distributions