Litimus scores rhythm, phrasing, and repetition the way an editor reads a manuscript — then shows its work sentence by sentence, so you can see exactly why it turned the color it did.
Edit your draft with the notes above, then paste the revised version back in — most writers see a real shift after a few targeted changes.
Everything below runs the moment you paste text — nothing leaves your session unless you save a report.
Plain text, a Google Doc link, or .docx / .pdf on paid plans. 25-word minimum for a stable read.
Sentence rhythm, phrase frequency, vocabulary spread, and repetition are scored independently, then combined.
A single likelihood score plus margin notes pointing at exactly which sentences moved it.
No black box score. Every signal below is visible in your report.
Human sentences swing between short and long; AI output tends toward a narrow, even rhythm. Low variance raises the score.
Phrases like "moreover," "it is important to note," and "in conclusion" cluster far more densely in generated text.
A low ratio of unique words to total words suggests safe, repeated word choices rather than a wide natural vocabulary.
Repeated three-word sequences across a passage are rare in human drafts and common in generated ones.
Overuse of em dashes and semicolons, or an absence of the small errors humans make, both shift the score.
Paragraphs of near-identical length and matching openers read as templated rather than composed.
We'd rather lose a sale than oversell what statistical detection can do. Every report on Litimus carries the same caveats we're stating here.
Every tier includes the full margin-notes report. Higher tiers add volume, file types, and integrations.
A one-time top-up for a single deadline — no subscription, no recurring charge. Adds 10,000 words to your daily limit, valid for 24 hours, stacks on top of any plan including Free.