AI Slop Detector: Does Your Writing Sound Like AI?

Look for patterns rather than single words: a cluster of words like delve, seamless and leverage, set phrases like “in today's fast-paced world”, em-dashes every other sentence, lists that always come in threes, and sentences that are all about the same length. One of these means little. Several together is the signature of unedited model output.

The usual suspects are delve, tapestry, testament, landscape, realm, seamless, robust, leverage, elevate, unlock, foster, pivotal, crucial and navigate. Phrases are stronger tells: “let's dive in”, “it's worth noting”, “plays a crucial role”, “a testament to” and “whether you're a beginner or an expert”.

No. Those tools run a statistical model and give a probability you can't inspect. This one checks for named, visible patterns and shows you every point it adds, so you can fix the writing instead of arguing with a number. A careful human can trip it and a well-prompted model can pass it.

Chat models use the em-dash far more often than most people type it, often several times per paragraph to stage a pause or a reveal. Plenty of good writers use them too, so the tool only scores density: above about one per 100 words it starts adding points.

AI slop is text generated in bulk with little editing: generic, padded, and built from the same stock moves, like the fake reversal (“it's not just a tool, it's a movement”) and emoji-bulleted lists. It reads fine at a glance and says almost nothing, which is why readers have learned to skim past it.

Cut the stock phrases and the first and last paragraph, which are usually throat-clearing. Replace abstract claims with a specific number, name or example only you would know. Vary sentence length on purpose, and say the point once instead of in three parallel items.

No. The score is computed by JavaScript in your browser tab; the text never leaves your device and nothing is logged.