FREE No signup · Runs in your browser
Paste it. Find the tells. Fix them before anyone sees.
The Slop Check reads your draft the way your most skeptical colleague does: it flags the vocabulary, the phrases, and the rhythm that make writing read like AI. Not a detector guessing who wrote it. A checklist of what gives it away, so you can fix it.
Your text never leaves this page. The check runs entirely in your browser: nothing is uploaded, nothing is stored, nothing is seen by anyone but you.
Tune the check to your voice
The check above runs the house rules. Voice Kit owners tune it: add the crutch words only you overuse, mute the rules that aren’t yours. Your tuning saves to your hub and loads every time you’re here.
Get the Voice Kit — $19Already own it? Log in to your hub and come back; the panel unlocks itself.
Your rules
Runs with every check on this page. The words below get flagged like the built-in vocabulary; unticked rules stay quiet.
The reference
What makes writing sound like AI?
Writing sounds like AI when it carries the tells: a stock vocabulary (“delve,” “robust,” “seamless”), stock phrases (“in today’s fast-paced world”), the punctuation tics, the “it’s not just X” construction, and sentences that all run the same length. None of these is proof on its own. Together, they read as machine-made in under three seconds.
Here is the full list the Slop Check runs, why each one reads as AI, and the fix. The tells all predate AI. Weak corporate writing used them for decades. AI just made them universal, which is why your reader now pattern-matches on them instantly.
i.Slop vocabulary
“Delve.” “Leverage.” “Robust.” “Seamless.” “Elevate.” “Harness.” “Tapestry.” “Holistic.” “Transformative.” Nobody says these words in a meeting, which is exactly why they read as generated. They are the words a model reaches for when it has nothing specific to say.
The fix: say the plain word. You don’t “leverage” a tool, you use it. Nothing “delves”; it looks.
ii.Stock phrases
“In today’s fast-paced world.” “At the end of the day.” “Now more than ever.” “It’s important to note.” “When it comes to.” “In conclusion.” These are throat-clearing. They carry zero information and your reader skips them, or worse, notices them.
The fix: delete the phrase. The sentence that remains is the sentence you meant.
iii.The em-dash tic
AI loves the long dash. One per paragraph can be earned. Three in two sentences is a signature, and by now most professionals recognize it on sight: “The draft was solid — mostly — until the ending.”
The fix: swap most of them for a comma, parentheses, or a period. Keep the one that earns its place.
iv.“It’s not just X”
“It’s not just a report, it’s a roadmap.” “This isn’t only about revenue.” “Not only X, but also Y.” The antithesis construction is the single most recognizable AI sentence shape of the decade. It sounds profound and commits to nothing.
The fix: say the Y. If the roadmap is the point, write about the roadmap.
v.Throat-clearing openers
Sentences that begin with “Moreover,” “Furthermore,” “Additionally,” “Notably,” “Ultimately.” Real people almost never open a sentence this way. Models do it constantly, because it fakes the connective tissue that actual thinking provides.
The fix: cut the opener. If the sentences don’t connect without it, the problem is the sentences.
vi.The rule-of-three pileup
“Clear, concise, and compelling.” “Faster, smarter, and more efficient.” One triad reads as rhetoric. A triad in every paragraph reads as a template filling itself in.
The fix: pick the one word that’s true and cut the other two.
vii.Treadmill rhythm
Every sentence the same length, the same shape, the same tempo. This is the most reliable tell of all, because it’s invisible word by word and unmistakable in aggregate. Human writing varies. A long sentence sets something up. A short one lands it.
The fix: read it aloud. Anywhere you fall into a rhythm, break it with a short sentence.
viii.Hedge stacking
“This could potentially help teams somewhat improve outcomes.” Models hedge because they’re built to avoid being wrong. A paragraph where nothing is asserted is a paragraph nobody needed.
The fix: decide what you actually think, then write that.
ix.Chatbot residue
Double asterisks around **key phrases**. Headers and bullet fragments in the middle of what should be an email. “I hope this helps!” These aren’t writing tells; they’re evidence. They mean the text was pasted straight from a chat window without a single read-through.
The fix: read it once before you send it. That’s the whole fix.
The problem was never using AI. Half your company is required to. The problem is shipping the unedited output and letting it speak for you. The tells are fixable. That’s the point of naming them.
Questions
What people ask
Is this an AI detector?
No. Detectors guess who wrote a text, and they guess wrong often enough to be dangerous. The Slop Check answers a more useful question: what in this draft reads as AI, and how do you fix it. It works the same whether a model wrote the draft or you did.
Should I just run it through a humanizer?
A humanizer rewrites AI text to slip past a detector. That hides the machine; it doesn’t make the words yours, and it bills you every month to keep doing it. If your writing sounds like AI, the fix is to put your voice in from the start. Here’s the full comparison, humanizer versus the Voice Kit, and why detection-evasion is the wrong job.
Where does my text go?
Nowhere. The check runs entirely inside your browser on this page. Your text is never uploaded, never stored, and never seen by us or anyone else.
What does it check for?
Nine tells: slop vocabulary, stock phrases, the em-dash tic, the “it’s not just X” construction, throat-clearing openers, rule-of-three pileups, treadmill sentence rhythm, hedge stacking, and chatbot residue like stray markdown. The full list, with fixes, is on this page.
It flagged something I wrote myself. Is it broken?
No, and that’s the uncomfortable part. The tells predate AI; corporate writing used them for decades, and models learned them from us. If the check flags your own sentence, the sentence reads as generated to your reader too. Cut it anyway.
Last updated July 2026. The tells list is refreshed as the models change.