Feature · Stage 3, Check the book

Ten times in two pages, and you cannot hear it any more.

An invented word stops registering as unusual to the person who invented it, which is exactly why it can pile up in one scene without you noticing. Your reader is meeting it for the fourth time on the page and has not stopped noticing at all.

Try it free
How it works

Measured against itself, not against a number.

1

A window of five hundred words

About two paperback pages. Ten mentions inside one is roughly once every two sentences, which is the point at which a word stops being vocabulary and starts being a tic. The window slides to wherever the mentions actually are, so nothing is missed by falling across a chapter break.

2

Compared with the word’s own rate

This is the part that makes it useful. A book’s main character is mentioned heavily everywhere, so a fixed threshold reports the title of the book: the first version of this flagged Alice in Alice in Wonderland. A passage only counts here when it runs at four times the rate that word keeps across the whole manuscript.

3

Your vocabulary, not everybody’s

Only words the scan read as invented. An ordinary English word falling ten times in a page is a fact about English, and a tool that reports it is a repetition checker, which you can get anywhere.

4

Nothing is called a mistake

A word repeats because the scene is about the thing it names, and some of these will be the climax doing its job. The report tells you where the passage is and how far off the book’s own rate it runs. The judgement stays yours.

Use cases

Where clustering shows up.

Fantasy and science fiction

The chapter that explains how the magic works, where the word for it lands nine times in a page.

Line editing

Finding the paragraphs to vary before a copy editor charges you to notice them.

Read-aloud and audiobook

Repetition that survives silent reading is unmissable when a narrator says it.

Late drafts

A scene rewritten three times often keeps a word from each version.

FAQ

Questions about term clustering.

Is this a repetition checker?

No, and the difference is the point. A repetition checker reads every word and has no idea which of them you made up. This reads only the vocabulary your book invented, and compares each word against the rate it keeps everywhere else, so it reports the passage that is unusual for that word rather than every word that appears twice.

Why does my main character not appear here?

Because a name mentioned heavily throughout the book is not clustering, it is the book. A passage has to run at four times the rate that term keeps across the whole manuscript before it is reported.

Is a cluster always worth fixing?

No. Nothing on this screen is counted as work anywhere in the product. Some of these are the scene where the thing finally matters, and the word belongs there. It is a place to look, not a list of errors.

Does it work on a very long manuscript?

A very large scan thins the mention positions it sends back, to keep the result a sensible size. This report is entirely a measurement of how close together the mentions are, so where that happens it says so and shows nothing rather than measuring a thinned list, which is evenly spread and would give believable figures that are wrong.

See which of your words is doing too much at once.

Free to preview, no sign-up required.