Lexical20%
A frequency-weighted dictionary of words and phrases that are fine alone but cluster suspiciously in AI text. Deterministic — no model involved.
delve · leverage · “a rich tapestry” · “navigate the complexities”
The free Slop Score
You lose interest the moment you realize it's AI.
So does everyone reading your work.
Paste anything. Get your Slop Score in seconds — a 0–100 reading of how AI-like it sounds, with the exact words and rhythms that give it away.
0 / 8,000500 more to go
Free · No signup · 500–8,000 characters
Try:
Three engines, one score
Each engine returns its own 0–100 read and flags spans in your text. The Slop Score is their weighted sum — higher means sloppier, on purpose.
Structural35%
Light NLP over rhythm and shape: tricolons, “it's not just X, it's Y”, uniform paragraph lengths, em-dash density, reflexive closers.
“It's not just a tool — it's a testament to…”
Epistemic45%
The one model-based read: does the text take a position, name specifics, and earn its certainty — or could it front any article on the topic?
Hedging without stance · zero proper nouns · both-sides closers
score = round(0.20 × lexical + 0.35 × structural + 0.45 × epistemic)
We show our work
A score is only worth trusting if it lands where it should. These reference texts are pinned in our test suite — every change to the engines has to keep them inside their band, or it doesn't ship.
Default-prompted AI marketing copy
raw GPT/Claude output
75–90
Corporate jargon memo
human, but slop-adjacent
45–60
Paul Graham essay excerpt
human
< 25
George Saunders essay excerpt
human
< 20
E.B. White, children's book
human
< 15
Scored in the wild
The latest public pages people ran through the score.
Get the rewriter when it launches
We'll email once. No newsletter, no nonsense.