Most teams don't have a strategy problem. They have a clarity problem.
Only the posts worth your time.
Built for LinkedIn.
Slop Mop reads ahead and folds away the junk, whether a person or an AI wrote it.
Chrome
Comet
Brave
Opera
All Chromium browsers
Microsoft Edgesoon
Safarisoon
Firefoxsoon
My feed had turned into something I scrolled past instead of read. Some of it was AI, plenty of it was people, and all of it cost attention. I built Slop Mop to read ahead for me, show its working, and let me overrule it whenever it's wrong.
Many posts too little worth reading
Generative AI made plausible writing nearly free, and feeds reward volume. People post more to be seen, tools write it for them, and the few posts worth your attention sit between the hooks, the bait and the tidy little parables.
- 01AI made it cheap. A post that took an hour now takes a prompt.
- 02Humans make it too. Broetry and engagement bait were here long before the models.
- 03Feeds reward it. Reactions are easy to farm, so the formula spreads.
- 04You pay in attention. Every post costs a few seconds to decide it wasn't worth it.
So Slop Mop reads ahead and does the deciding. Here's the kit it uses.
Nine tells of slop
The model asks nine separate questions about the writing. Six come from Graphite's research on AI tells; the last three are feed habits we added. One spike isn't enough: slop usually shows several at once.
-
Scroll-stopper
A dramatic one-line opener, a gap, then one-line paragraphs walking to a moral. Built to stop the thumb, not to say anything.
e.g. I got rejected 47 times. / Then everything changed. -
Engagement bait
Asks for the reaction instead of earning it: comment a keyword, repost to help your network, follow for part two.
e.g. Comment “GUIDE” and I'll DM you the template. -
Hype words
Marketing vocabulary standing in for a claim. Big verbs, no numbers.
e.g. This game-changing framework will unlock unprecedented growth. -
Empty praise
Calls something meaningful or profound and never says what it did.
e.g. An incredibly meaningful, truly transformative experience. -
No-catch promises
All benefit, no cost. Real work has trade-offs; this writing never does.
e.g. Scale your output without sacrificing quality or hiring. -
“Not X, but Y”
Defines things by what they aren't, over and over, so every line sounds like a revelation.
e.g. Leadership isn't a title. It's a mindset. -
Flowery prose
Flourish in place of the plain claim.
e.g. Growth is a delicate dance between vision and humility. -
Stiff phrasing
Stock transitions and hedges that make a sentence longer and say less.
e.g. Furthermore, this approach may provide additional benefits. -
Too-tidy story
An anecdote that exists only to land exactly on its lesson.
e.g. A junior employee once asked me a question I'll never forget.
A dramatic one-line opener, a gap, then one-line paragraphs walking to a moral. Built to stop the thumb, not to say anything.
The value of humanity, included
Useful beats authentic. Three counter-signals sit beside every verdict, and they can only pull a score down. A useful post survives the mop however it was written.
One particular person, not anyone
Specific detail from the writer's own life or work, phrasing that's clearly theirs, uneven rhythm. “I” and “my” alone don't count. It takes a little off the score.
Something you could use today
Facts, numbers, a named example, a step to follow. Who or what wrote it doesn't matter: useful AI-assisted writing scores as well as useful writing typed by hand.
What readers actually did
Comments and reposts count for far more than one-tap reactions, and hundreds of reactions with no discussion is discounted as hollow. The other half of the shield.
Judges the writing. Never the writer.
Slop Mop has no opinion about you, and none about the person who wrote the post. It never learns who wrote a post, and tells nobody what it thought.
- No names go out with a post. Not yours, not the author's. Only the post text and its public counts are sent to be judged.
- Posts aren't stored, unless it got one wrong. The server keeps a one-way hash so a post isn't judged twice. Only when your vote disagrees with its verdict is the text kept, so a person can review it and the scoring can learn.
- No profiles, of anyone. No score is ever attached to a person, and no browsing history or LinkedIn profile data is collected.
- Nothing is reported. No author is flagged, muted or told. A verdict only changes what you see, in your own browser.
- Your votes can't be traced to you. They're kept apart from your account, never shown to anyone or sent to the author, and only ever counted in aggregate.
Reads ahead, so you don't have to
Seven steps between a post loading and you deciding it wasn't worth it. All but the last finish before you scroll there.
Read the full breakdownPosts are read before they reach your screen.
Each check loads the sign library: nine tells, three counter-signals.
An AI model built for calibrated decisions answers twelve questions at once.
The verdict lands on the mop beside the post.
Past your threshold, the post folds into a strip you can reopen.
The score, the strongest signs, the Slopprint and the three bars.
Not sure, No, Maybe or Probably. Your vote overrides the model in your feed.
One mop for every feed
Slop Mop works on LinkedIn today. The same mop, fold strip and details panel are coming to the other feeds you scroll, the articles and websites you open, and the email you triage.
social feeds
articles & websites
browsers
comet, brave, opera, arc, vivaldi and other chromium browsers install from the chrome web store
One feed is free. The XL mop does the whole house.
Free cleans the one feed you pick, with 250 checks a day, for good. Slop Mop XL takes the mop everywhere you read, and it's coming soon.
- The one feed you pick, read ahead and mopped (LinkedIn today)
- Hide or Highlight, at three sensitivities
- The full details panel and the Slopprint on every post
- Your vote overrides the model in your feed
- 250 checks a day
Best value save 17% on yearly · or $5.99 a month
- Everything in Slop Mop
- Every feed: LinkedIn, X, Reddit and Facebook
- The sites you read: articles, comments and reviews
- Your inbox, with Slop Mop for Gmail and Outlook
- More checks every day than the free 250
- Up to five browsers on one account
Questions? We've mopped up a few.
What is Slop Mop?
Slop Mop is a browser extension, free for one feed, that reads the posts in your feed before you reach them, judges the writing, and folds away what reads like slop. Nothing is deleted: a folded post reopens with one click, and every verdict shows the signs behind it.
Today it works on LinkedIn in desktop Chrome and Chromium browsers such as Comet, Brave and Opera. Microsoft Edge, Safari and Firefox are coming soon. X, Reddit, Facebook, Substack, Medium, email and the rest of the web are coming.
Is Slop Mop an AI detector?
No. Slop Mop is a bad-writing detector. AI-assisted writing can be excellent and human writing can be spectacular slop, so it asks a more useful question: is this worth your attention?
How does it decide whether a post is slop?
An AI model designed for calibrated decisions answers twelve questions about each post in parallel: nine tells of slop, two counter-signals (does it sound like a person, is it useful to readers) and how likely it is to be human-typed. Reader response, counted from reactions, comments and reposts, is the third counter-signal.
The answers become one score out of 100. Counter-signals can only lower it. How the score adds up.
Does it block, delete, mute or report posts?
No. Slop Mop only changes what you see, locally in your browser. Hide folds a post into a strip you can reopen; Highlight leaves it visible and tints its mop icon yellow or red. Nothing is sent to LinkedIn and no author is ever reported.
What happens when it gets one wrong?
Click the mop beside the post and vote: No, Maybe or Probably, or Not sure to clear your vote. Your vote overrides the score in your feed straight away, with or without an account. It also counts toward the community vote on that post: in full once you have an account, and as a hundredth of a vote while you're trying Slop Mop without one.
When your vote disagrees with Slop Mop's verdict, the post's text is saved and a person at Artimovo reviews it; the posts it got wrong become training examples for tuning the tells and the score. Nothing that leads to you or the author is saved with it: it judges writing, not people. You can't vote on your own posts: the vote panel doesn't appear on them. It will be wrong sometimes. That's why it shows its working and gives you the last word.
Is it free?
Slop Mop is free for one feed, with 250 checks a day, for good. LinkedIn is the first feed it reads; X, Reddit and Facebook are coming soon. Slop Mop XL, coming soon, takes the mop to every feed, the sites you read and your inbox, for $5.99 a month or $59.99 a year. It renews automatically until you cancel, and you can cancel any time.
Do I need an account to vote?
No. You can vote from the first post you see, and your vote changes your own feed straight away. Without an account, each vote counts as a hundredth of a community vote, so a hundred trial votes add up to one. Free and XL accounts count the same.
Signing up makes every vote that browser has already cast count in full, so the votes you made while trying Slop Mop stick. Votes still can't be traced to anyone: the server stores how much each vote counts, never which account cast it. If you've used Slop Mop since its early versions (0.x), every vote you've cast counts in full.
What leaves my browser?
The post text and its public engagement counts go to api.slopmop.lol to be judged. Your name and the author's name are never sent. The server keeps a one-way hash of the text, not the text, unless your vote disagrees with its verdict: then that post's text is kept for review and training. The privacy policy lists everything it stores.
Does Slop Mop read my email?
Only with Slop Mop XL, through Slop Mop for Gmail and Slop Mop for Outlook, and only the emails you open or ask it to check. Personal details are taken out in your mailbox before anything is sent: names, email and street addresses, phone numbers, links, card and bank numbers, ID numbers and dates of birth become placeholders like [NAME]. We take every step we can to catch names written in the body of an email, but can't guarantee every one. Only what the sender wrote is sent, never attachments, signatures or the earlier messages quoted below. Spam, private, legal, health and transactional mail is skipped unless you check it by hand. Email text is never stored, voting isn't available for email, and email is never used for training. The privacy policy has the details.
Can Slop Mop see who I email?
No. Slop Mop never sees who an email is from or who it's to: the sender and recipient names and addresses are taken out in your mailbox before anything is sent. Slop Mop judges the writing in an email, never who sent it or who it's for.
Is Slop Mop open source?
The original version of the extension and judging server is MIT licensed and public on GitHub. Everything about it, and the link to the code, is on the open-source page.
About a minute to install. No account, no API key, nothing to configure before it starts reading ahead.