Sloppenheimer ✨ Powered by AI, obviously Detect slop ✨
🏆 #1 on Product Hunt (we asked ChatGPT to say that)

Is it slop?
Great question.

Sloppenheimer uses cutting-edge AI to detect cutting-edge AI. It's not a tool — it's a paradigm shift. Screenshot it, paste it, upload it, and let a large language model decide whether a large language model wrote it. What could go wrong. Let's dive in. 🚀

Start detecting ✨ Watch demo (there is no demo)
🔒 Your key never leaves your tab 📱 Works on your phone, like a person 🧠 Trained on nothing
A woman gazing at a glowing neon brain surrounded by iridescent bubbles
Hero image · as the brief demandedSloppenheimer verdict: 96/100 · Pure, uncut slop. "This is what happens when an AI tries to paint a dream and ends up inventing a shampoo commercial for neural networks." She is thinking about efficiencies.
no key A house key is baked in, so just go. ✨ Or bring your own: it lives in this tab's session storage and dies when you close it. We are a static HTML page. We couldn't steal it if we tried, and we can't be bothered to try. Get an Anthropic key · or an OpenAI one.
Paste a screenshot here ⌘/Ctrl + V, drop an image, or tap to choose one On a phone: long-press this box → Paste. Or tap to pick from your photos.
It's not a bug — it's a feature.  ✦  Great question!  ✦  I hope this email finds you well.  ✦  Certainly!  ✦  You're absolutely right.  ✦  In today's fast-paced world.  ✦  Let's delve in.  ✦  A rich tapestry.  ✦  It's important to note.  ✦  Apologies for the confusion.  ✦  As an AI language model.  ✦  Game-changer.  ✦  Thoughts? 👇  ✦  I've 10x'd my workflow.  ✦  Agree?  ✦ 
A purple fantasy landscape with a panda in a boat, a unicorn, a dragon, an owl and a rubber duck
Exhibit A · slop, figurativeSloppenheimer verdict: 96/100 · Pure, uncut slop. "This is what happens when you feed a neural network every Lisa Frank folder and a Tolkien wiki, then ask it to hallucinate on maximum sparkle mode." Note the rubber duck. Nobody asked for the duck.
10M+things analysed*
94.7%accuracy†
40 hrssaved per week‡
0jobs lost§

* number generated by AI   † see "accuracy"   ‡ by a man on LinkedIn, allegedly   § so far

Trusted by absolutely nobody at

Synergiadelve.aiTapestry LabsSeamless Robust SolutionsParadigm Shift Inc.Grindset CapitalYour Line Manager

How it works. Seamlessly.

Three simple steps. We tried to make it two but the third one tested better.

1

Upload a thing

A screenshot. A LinkedIn post. A photo of a woman laughing at a salad with eleven fingers. Anything you suspect was extruded rather than made.

[IMAGE PLACEHOLDER]prompt: "person uploading a screenshot, ultra realistic, 8k, trending on artstation, hands hidden for safety"
2

We ask a robot

Your evidence is sent to a large language model, which is asked, in effect, "did one of your lot do this?" It looks for the tells. Em-dashes. Tricolons. The word "delve". The seventh finger.

[IMAGE PLACEHOLDER]prompt: "a friendly robot in a lab coat inspecting a photograph with a magnifying glass, Pixar style, no text, NO TEXT, why is there text"
3

The robot says "Great question"

And then, after apologising for any confusion, gives you a score out of 100, the evidence, and a verdict you can screenshot and post underneath the thing. Circle of life.

[IMAGE PLACEHOLDER]prompt: "a golden trophy that says CERTIFIED SLOP, dramatic lighting, hyperdetailed, the trophy text says CERTIFED SLPO"

What our users are saying.

Real testimonials from real people. We've verified this by asking them, and they said yes, which is what a real person would say.

★★★★★

"I used to spend 40 hours a week spotting AI content. Now I spend 40 hours a week telling people I've 10x'd my workflow."

C
Craig · Head of Efficiency (a title he gave himself)
★★★★★

"I could tell it was AI from the very first word. The first word was 'The'."

T
Tamsin · Professional Noticer
★★★★★

"AI is coming for my job. I've decided to get there first and do it badly myself."

D
Dev · Copywriter (for now)
★★★★★

"As an AI language model, I do not have personal experiences, but I hope this testimonial finds you well."

G
GPT-4o · Satisfied Customer
★★★★★

"It's not a tool — it's a movement. A movement — not a tool. Not a tool, a movement."

S
Our marketing team · who have read one thread about tricolons
★★★★☆

"I make everything by hand, from scratch, with feeling. This app said my poem was 91% slop. I have concerns about the app."

R
Rowan · Real Artist

The tells we delve for.

Every model has a tell. So does every person who says they don't use one.

— the em-dash —delvetapestrya testament toit's not X, it's Yrobust, seamless, scalableCertainly!Great question!You're absolutely rightin today's fast-paced worldit's important to noteI hope this finds you wellnavigate the landscapeunlock · leverage · elevatethe seventh fingertext that says "TEXT"waxy skin, perfect teethemoji bullets 🚀Thoughts? 👇every paragraph the same lengtha conclusion that says "in conclusion"Agree?
A pile of pink slime on a table
Exhibit B · slop, literalSloppenheimer verdict: 10/100 · Disgustingly human. "If this is AI-generated, then the singularity will be heralded by a puddle of bubblegum and regret." It is AI-generated. Grok made it. The fox is very well-read but it has never seen slime. We're leaving this in.

Simple, transparent pricing.

No hidden fees. No fees. No company. Bring your own key and we'll bring the judgement.

Free

£0/forever
  • Everything
  • Bring your own key
  • Unlimited slop
  • Judgement included
Get started

Pro

£0/month
  • Exactly the same
  • But you feel better about it
  • A badge that says "Pro"
  • Priority access to the same thing
Go Pro (nothing happens)

Enterprise

Contact sales
  • There is no sales
  • We are not a company
  • Do not contact us
  • SSO (Someone's Screenshot, Obviously)
Do not contact us

Frequently asked questions.

Nobody has asked these. We generated them. It felt right.

Does it actually work?

Certainly! Great question. Yes, mostly. It asks a frontier vision model to look for the fingerprints generators leave behind and to weigh them against evidence of a human. It's a bit like asking a fox to audit the henhouse, but the fox is extremely well-read and has, frankly, seen a lot of hens. Short generic text is nearly undecidable and the app will say so instead of pretending.

Is my data safe?

Your key sits in your browser's session storage and is deleted the moment you close the tab. Your screenshot goes straight from your browser to whichever AI company issued the key, who pinky-promise. We see nothing. We are a static HTML page on a man's VPS. We couldn't spy on you if we wanted to, and we don't have the bandwidth.

Will AI take my job?

Not if you keep posting like that.

Isn't using AI to detect AI a bit…?

Yes.

I'm a real artist and I find this offensive.

Noted. Lovely brushwork.

I've saved 40 hours a week with AI. Where do I post about it?

You've already posted about it. Twice. We ran it through the detector. You're not going to like the number.

Can I get it on my phone?

Yes. On iPhone: Share → Add to Home Screen. On Android: the browser will offer to install it, or use the menu. It's a PWA, which stands for something.

How do I screenshot a bit of my screen?

You don't have to. Tap Screenshot in the app, pick the window or tab, then drag a box around the suspect. Works on laptops. On a phone, press the side button and volume up together, like a person, then paste it in.