Paste a scene or chapter and get instant prose statistics: dialogue ratio, adverb density, filter words, sentence rhythm, and readability. No signup. No upload.
Everything runs locally in your browser. Nothing is stored, sent, or used for AI training.
Each bar is one sentence (first 120 shown). Healthy prose varies — long flowing sentences broken by short punchy ones. A flat wall of same-height bars reads monotonous.
Words that weaken prose when overused: very, really, just, suddenly, somehow, quite, almost, started to, began to. Each is worth a second look — most can be cut or replaced with a stronger verb.
Words like saw, heard, felt, realized, noticed, watched put a lens between the reader and the scene. “She felt cold air on her neck” is weaker than “Cold air brushed her neck.”
Your top non-trivial words. Repetition is invisible while drafting — if a distinctive word appears every few hundred words, readers notice.
These are the mechanics. ProseEngine scores the story — tension, pacing, hook power, emotional impact, and 10 more metrics on every scene, plus AI that fixes what it finds.
Try ProseEngine FreeDialogue ratio. The share of your text inside quotation marks. Most commercial fiction sits roughly between 30% and 60%, but genre matters: thrillers and romance lean talky, epic fantasy and literary fiction lean descriptive. The most useful comparison is your own scenes against each other — an action scene at 15% and a confrontation at 70% is fine; every scene at 45% is monotone.
Adverb density. A common editing guideline is fewer than ~10 -ly adverbs per 1,000 words. Adverbs aren't banned — but “walked slowly” is almost always weaker than “trudged.”
Filter words. Saw, heard, felt, noticed, realized, watched, seemed — these report perception instead of rendering it. Cutting them is the fastest mechanical route to deeper point of view.
Sentence length. Average commercial fiction runs about 12–18 words per sentence. The average matters less than the variation: rhythm comes from contrast.
Passive voice. An estimate based on “was/were + past participle” patterns. Some passive is natural; pages of it drain momentum.
Readability. Flesch Reading Ease — most popular fiction scores 70–90 (easy to very easy). Lower isn't wrong, it's just denser; know which you're writing.
Standard ranges agents and publishers expect for debut novels.
| Genre | Typical word count |
|---|---|
| Epic / High Fantasy | 90,000 – 120,000 |
| Science Fiction | 85,000 – 110,000 |
| Thriller / Crime | 70,000 – 90,000 |
| Mystery | 70,000 – 90,000 |
| Romance | 70,000 – 90,000 |
| Literary Fiction | 70,000 – 100,000 |
| Young Adult | 55,000 – 80,000 |
| Middle Grade | 25,000 – 55,000 |
| Novella | 17,500 – 40,000 |
No. The analyzer runs entirely in your browser with plain JavaScript. Your text is never sent to a server, never stored, and never used for AI training. You can disconnect from the internet after the page loads and it still works.
Word count and reading time, dialogue ratio, adverb density, filter words, crutch words, a passive-voice estimate, sentence-length rhythm, your most-used words, and Flesch readability.
Roughly 30–60% for most commercial fiction, but it varies by genre and scene type. Compare your own scenes against each other rather than chasing an absolute number.
A common guideline is under ~10 -ly adverbs per 1,000 words. Treat each one as a prompt to consider a stronger verb, not as a rule.
No — and be suspicious of any word-counter that claims it can. These are mechanical signals: useful for editing passes, blind to story. Tension, pacing, character depth, and emotional impact need deeper analysis — that's what ProseEngine's 14-metric scene scoring does.