AI Slop Score
Score a draft on the tells that give AI writing away, then take the rewrite. No model call, no account, nothing leaves your browser.
What is an AI slop score?
A count of the stylistic habits that mark writing as machine generated: em dashes, stock phrases like delve and a testament to, lists of exactly three, emoji bullet lines, the it is not just X it is Y construction, and sentences that are all the same length. Higher score, more obvious.
Your draft
0 words, 0 characters.
Score
Paste a draft to score it.
We hold ourselves to this
Em dashes and en dashes are banned across every page on this site by an internal writing rule, for exactly the reason this tool flags them. That rule predates the tool. If you find one in our copy, it is a bug. The two dash characters in the source of this page exist only so the checker can match them.
Why this matters for X payouts
X Original Content Rewards, which replaces Creator Revenue Sharing from 2026-09-08, pays on qualified impressions from paying Premium users and explicitly excludes reposted and reproduced content. The program is built around original work. Writing that reads generated and interchangeable is arguing against your own case.
Originality is a payout gate
Reposted or reproduced content does not qualify. Generic model output is the easiest kind of content to reproduce.
The gates are steep already
500 verified followers and 500,000 verified Home Timeline impressions in the prior 90 days, replies excluded. Do not waste impressions on posts nobody finishes.
Readers pattern match fast
A regular X reader spots emoji bullets and an em dash in the first second. The scroll happens before the argument does.
Every tell this tool checks
Em dashes and en dashes
auto fixedThe single most reliable tell. Almost nobody types an em dash on a phone keyboard, and language models produce them constantly.
Stock model phrasing
auto fixedPhrases like delve, a testament to and in the ever-evolving landscape appear in model output at rates far above normal writing.
The it is not just X, it is Y construction
auto fixedThis antithesis template is the most recognizable sentence shape in model output. Once you notice it you cannot unsee it.
Triadic lists
manualThree items joined with a final and, over and over. Models default to the rule of three whether or not there are three real things.
Emoji bullet lines
auto fixedLines opening with a checkmark, rocket or fire emoji are a formatting habit of chat models, not of people posting from the app.
Uniform sentence length
manualHuman writing varies hard: a fourteen word sentence next to a three word one. Model output settles into a narrow band and reads flat.
Rhetorical question then answer
auto fixedThe Why? Because pattern, and its cousin The answer is simple, are filler that models use to build rhythm.
Markdown that X will not render
auto fixedDouble asterisk bold and heading hashes survive a copy paste from a chat window. X renders them as literal characters in a post.
Hashtag stuffing
auto fixedMore than two hashtags in one post is a model habit picked up from marketing copy. It does nothing for reach on X.
Write with AI without sounding like it
OpenTweet learns your voice from posts you have already published and writes in it, instead of defaulting to model house style. Then it schedules the result, recycles what worked, and publishes through the API or an AI agent.
7-day free trial, then $11.99/mo. Cancel anytime.
AI Slop Score FAQ
Why are em dashes such a strong tell?
Because almost nobody types one. There is no em dash key on a phone keyboard, so a person writing a post reaches for a comma or a period. Language models produce em dashes constantly. It is the highest signal, lowest effort thing to fix.
Is this an AI detector?
No, and you should not trust anything that claims to be one. This scores stylistic habits that correlate with model output. Human writing can score high and careful model output can score low. It tells you how the post reads, not who wrote it.
Why does uniform sentence length matter?
Human writing varies hard, a fourteen word sentence next to a three word one. Model output settles into a narrow band. The tool measures the coefficient of variation of your sentence lengths and flags anything under 0.32 across four or more sentences.
Does AI-sounding writing affect X Original Content Rewards?
Reposted and reproduced content does not qualify for payout under Original Content Rewards. The program is built around original work, so writing that reads generated and interchangeable is working against the thing you are trying to get paid for.
Does the rewrite change my meaning?
It tries not to. Every edit is a deletion, a punctuation swap or a one-phrase substitution, never a rewritten sentence: dashes become commas or periods, filler is deleted, phrases like delve into become read, and the not just X it is Y construction keeps only the second half. Read it before you post it.
Do you follow this rule yourselves?
Yes. Em dashes and en dashes are banned across every page of this site by an internal writing rule, for exactly the reason this tool flags them. That rule predates this tool.
Related free tools
Engagement Bait Checker
Flag the solicitation phrases that cost you the rewards program.
AI Tweet Rewriter
Rewrite a post toward a specific goal, not just a cleaner one.
Character Counter
Check the rewrite still fits before you post it.
X Analytics CSV Analyzer
Find out which of your posts actually landed.
See also all free tools, features, and pricing.