Every novelist has crutch words. "Just," "really," "somehow", and the surprise nouns and verbs you don't notice you keep reaching for. This tool surfaces your top 20 most-used words, flags weak filler, counts -ly adverbs, and scores your vocabulary diversity. Built for fiction and long-form self-editing.
Why this exists
Every writer has pet words: phrases or words that creep into the prose without you noticing. They're invisible when you're drafting and obvious to a reader. The fastest way to find them is to count.
This tool does that one job: it counts. The top 20 non-stopword words from your text, ranked by frequency, with weak/overused words flagged in red. No autoplay grammar suggestions, no AI rewrite buttons, no trial paywall.
How to use it for self-editing
- Run a chapter at a time. 5–10K words is the sweet spot. Whole-book frequency drowns out chapter-level patterns.
- Look at the top 10 first. If a word at position 1–3 surprises you, that's a pet word. Common culprits: "look", "smile", "nod", "back", "eyes", "head", "hand", "just".
- Check the watch-words panel. Anything appearing more than 3–4 times in a chapter is worth interrogating. "Really" appearing 8 times almost always means at least 5 are doing nothing.
- Check the adverbs panel. If "suddenly" or "quickly" or "softly" tops the list, the corresponding verbs are probably under-working.
- Don't fix everything. Aim to cut 30–50% of any over-used word, not all of it. Repetition used deliberately is fine; the goal is to remove accidental repetition.
What "stopwords" means
By default the tool filters out the ~150 most common English words ("the", "and", "of", "is", "to" etc.). These dominate any frequency count and crowd out the words that actually carry meaning. The toggle above the panels lets you include them if you want to see total frequency instead.
Stopwords are the standard list from NLTK, the open-source natural-language toolkit. Public domain.
The watch list
The flagged words are the standard self-edit list common across writing manuals (Strunk & White, Sol Stein, Renni Browne & Dave King's Self-Editing for Fiction Writers):
- Intensifiers: really, very, definitely, totally, completely, absolutely, perfectly, simply
- Hedges: basically, literally, actually, kind of, sort of, rather, quite, almost
- Vague nouns: thing, things, stuff, something, nothing
- Cliché motion: suddenly, somehow
- Filler: just
These aren't forbidden words. Used once or twice they're fine. Surfacing the count is the point. A writer who sees "really" 14 times in a chapter usually thinks "oh."
Vocabulary diversity, in plain terms
The diversity score is unique meaningful words divided by total meaningful words (after stopwords removed). It's a rough measure of how varied your vocabulary is in this specific text.
- 60%+: high variety. Literary fiction, dense exposition, technical writing.
- 40–60%: healthy for most fiction. Conversational but varied.
- 25–40%: noticeable repetition. Common in dialogue-heavy scenes or simple voices.
- <25%: heavy repetition. Either a deliberate stylistic choice (a child narrator, a tight POV with limited vocabulary) or a sign the prose is leaning on too few words.
The score drops naturally as sample size grows. A 100K-word novel will score lower than a 1K-word excerpt of the same writing because vocabulary plateau is real. Use the score to compare similar lengths against each other, not as an absolute target.
Beyond word repetition
Word repetition is one slice of what makes prose work or not. Sentence-level patterns matter too: sentence length variety, paragraph rhythm, dialogue density. The readability checker covers some of that.
And neither tool tells you whether you actually wrote today, this week, this month. That's what Authorlytica does: log a session in 10 seconds, see streaks, charts, pace projections, year-in-review. Free forever plan.