Count how often each word, phrase and character appears, with case, punctuation and stop words as switches rather than silent assumptions.
Waiting for text
Counts words, repeated phrases and characters. Every counting decision is a switch above.
The counts appear here
Every word with its share, every repeated phrase, and every character with its code point.
Word Frequency Analyzer runs entirely on your device — nothing is uploaded.
You’re on the Free plan. Compare plans
Three counts over one paste, and between them they answer one question: what is this text actually made of. Words find the subject. Repeated phrases find the boilerplate — the sentence that appears in every record, the disclaimer pasted forty times. Characters find the one stray glyph that is breaking something downstream.
Counting words requires decisions, and here each one is a switch rather than an assumption. Is Dog the same word as dog? Is dog with a comma the same as dog? Does the count at all? Every frequency tool answers these; most answer them silently, which is why two tools give different numbers for the same text and neither explains itself.
Stop words are off by default. The being the most common word in English prose is a true and useful fact about the text, and a tool that removes it before being asked has quietly edited your data. Turn the switch on when you want the subject matter rather than the grammar — for keyword work that is usually what you want, and it should still be your choice.
Diversity is unique words over total words. It says how repetitive the text is and nothing whatever about how good it is. A low number means words repeat, which is expected in a technical document and unexpected in a short story. Alongside it is the count of words appearing exactly once, which is the other half of the same picture.
Phrases never cross a line break, because the last word of one paragraph and the first word of the next are not a phrase. Stop words are always kept when counting phrases, since removing them would join words that were never adjacent and invent phrases that are not in the text. Nothing is uploaded: the text stays in the browser.
Because removing them edits your data before you asked. The being the most common word in English prose is a real fact about the text. The switch is there when you want the subject rather than the grammar.
Unique words divided by total words. It measures repetition and nothing else — it is not a readability score and says nothing about quality. Low is normal for technical writing and notable in a narrative.
Because the last word of one paragraph and the first of the next are not a phrase. Letting phrases cross the break would invent phrases that do not occur in the text.
Removing them would make words adjacent that were never adjacent, producing phrases the text does not contain. For phrases, the function words are part of what is being counted.
Almost always because of case, punctuation handling or stop words. Those three are explicit switches here, so the number on screen is one you chose and can reproduce.
No. All three counts run in your browser, and no part of the text reaches a server or analytics.