AI Text Slop Detector
Paste any draft - a README, email, landing page or blog post - and this tool scans it against common signs of AI-generated writing: self-reference phrases, collaborative filler ('let me know if'), not-just-X-but-Y contrast constructions, staged openers, summary clichés, overused buzzwords like 'robust' and 'seamless', excessive em-dashes, rule-of-three lists and unsupported authority claims. Matches are highlighted inline with a severity score, so you can revise a draft to read more like a human wrote it.
Reviewed by Vincenzo Manto, Founder & Lead Developer · Last updated October 12, 2026
How it works
- 1
Paste or type your draft into the text box.
- 2
The tool scans it locally against a set of known AI-writing patterns.
- 3
Flagged phrases are highlighted by severity, with a 0-100 slop score and category breakdown.
Why AI writing patterns matter
As more content online is AI-assisted, readers and editors have gotten sharper at spotting the tells: hedged contrasts, collaborative sign-offs, and a narrow set of buzzwords that large language models reach for by default. Catching these patterns before publishing helps a draft read as the writer's own voice rather than a generic AI output, which matters for trust, SEO content guidelines, and editorial review.
How the detection works
The tool checks text against a curated list of regex patterns grouped by category and severity, from critical self-reference phrases down to low-weight stylistic signals like em-dash frequency. Each match is highlighted inline and contributes to a 0-100 score, capped and labeled as likely human, mixed, or likely AI slop. Because it is pure pattern matching with no machine learning model, it runs instantly and entirely client-side.
Frequently asked questions
Does this prove the text was written by AI?
No. It flags stylistic patterns that are common in AI-generated text, such as filler phrases and clichés. A human can write these too, and AI text can avoid them. Treat the score as a signal for revision, not a verdict.
Is my text sent to a server?
No. All analysis runs locally in your browser using regex pattern matching. Nothing is uploaded or stored.
Which languages does it support?
English and Italian patterns are both checked automatically, since any pasted text may be in either language.
What does a high score mean?
A high score means the text contains many phrases associated with AI writing, such as 'as an AI language model', 'it's not just X, but Y', or buzzwords like 'robust' and 'seamless'. Rewriting flagged phrases in your own words will lower the score.
Why are em-dashes flagged?
AI models tend to overuse em-dashes compared to typical human writing. A few are normal; many in a short text is a common tell.