Free Tool — Client-Side Analysis

Check Any Text for AI Patterns

Paste text below to score it across four statistical signals. Everything runs in your browser — no text leaves your device, no account required.

AI Text Detector● Client-Side — No Data Sent
AI Phrase Matches
Statistical heuristics only. For high-stakes decisions, use a dedicated detector from our benchmark. Curious what 0% looks like? Try pasting from Chapter 3 of the Poignant Guide. See the Human Writing Standard →

How the Signals Work

{}
Input Text
Tokenized into words and sentences in your browser.
Phrase Scan
Checked against 22 AI-associated transition phrases. Density scored.
TTR + Variance
Type-token ratio and sentence length CV computed.
AI Score
Weighted composite: phrase density 45%, vocab 30%, variance 25%.

Frequently Asked Questions

Is this tool accurate?

It is a statistical heuristic, not a machine learning model. For production use or high-stakes decisions, use a dedicated detector from our benchmark — Originality.ai (91% accuracy) or GPTZero (87%) are the strongest options.

Does this send my text anywhere?

No. All computation runs entirely in your browser. No text is transmitted to any server. You can verify this in your browser network tab while running an analysis.

Why does humanized AI text score low?

AI humanizer tools specifically remove the patterns this detector measures. Even dedicated ML detectors struggle with humanized text — in our bypass study, accuracy dropped an average of 31 percentage points across all tested tools.

Why does _why's Guide score near 0%?

_why's writing has wildly variable sentence lengths (burstiness CV 0.94 vs AI average 0.31), zero formulaic transition phrases, and high vocabulary diversity (TTR 0.74 vs AI average 0.57). It is statistically the opposite of AI-generated text. Full analysis here.

Related Resources