AI Detector — Instant Heuristic Read on Any Text
Who this is for: Anyone who wants a quick second opinion on a draft before submitting or publishing — and who wants the signals spelled out, not just a percentage.
Instant heuristic read on whether text reads as AI-written or human-written. Runs locally in your browser — nothing is uploaded.
Quick answer: Paste text and get an instant heuristic score based on sentence-rhythm variance, burstiness and stock-phrase density. It runs locally in your browser, nothing is uploaded, and every signal is explained. It is a reference read, not proof.
| Cost | Free, unlimited |
|---|---|
| Where it runs | 100% in your browser |
| Method | Burstiness + stock-phrase heuristics |
| Uploads | None — local math only |
A heuristic read, honestly labeled
This is a statistical reading, not a verdict. It measures the signals that AI detection research has focused on: burstiness (how much sentence length varies — humans vary a lot), repetition of stock phrases and transitions, uniformity of sentence openings, and vocabulary predictability. You get a score plus the specific signals found, so you can judge the text yourself.
Because it runs entirely in your browser — the analysis is plain math on the text you paste — nothing leaves your device, and results appear instantly with no usage cap.
No detector, including paid ones, reliably identifies AI text. False positives against wholly human writing are documented, and skilled human writers who write in steady, uniform rhythm can score as machine-like. Treat this tool as one input among several, never as proof.
Common uses
- Check whether your edited draft still reads machine-like
- Compare two versions of the same passage before publishing
- Sanity-check a guest post or freelancer submission
- See which specific patterns push text toward the machine pole
- Test whether your humanize passes actually changed the statistics
The signals behind the score
Burstiness: human writing mixes 4-word sentences with 30-word sentences; machine text drifts toward the middle. Stock phrases: certain transitions and openers appear far more often in AI drafts (the tool checks a curated list and its frequency). Opening uniformity: sentences that start the same way, paragraph after paragraph, push the score toward machine. Lexical predictability: common-word ratios that flatten out in generated text.
What a high machine score does and does not mean
It means the statistical patterns resemble machine-generated text. It does not mean the text was generated by a machine — heavily edited AI text, non-native English writing with formal training, and bureaucratic prose can all score high while being legitimate. Use the score to decide where to revise, not to accuse.
Frequently Asked Questions
Is this AI detector accurate?
Does my text get uploaded anywhere?
Why does wholly human writing sometimes score as machine-like?
Can I use this score to prove text is AI-written?
More guides
What 'free' actually means across AI humanizer tools — word caps, daily limits, and the no-account difference.
More writing tools
Rewrite any text in three modes — humanize, rewrite, or paraphrase. Free 5 passes a day, 300 words each, no signup.
A free alternative to QuillBot's AI humanizer: 300 words per pass and 5 passes daily without an account, vs a ~125-word cap on the free tier of the brand tool.
Restate any passage in genuinely different wording and sentence structures — same ideas, no copied phrasing. Free, no signup.